WO2016155372A1

WO2016155372A1 - Method and device for predicting air quality index

Info

Publication number: WO2016155372A1
Application number: PCT/CN2015/098979
Authority: WO
Inventors: 张震
Original assignee: 北京奇虎科技有限公司; 奇智软件（北京）有限公司
Priority date: 2015-03-27
Filing date: 2015-12-25
Publication date: 2016-10-06
Also published as: CN104751242A

Abstract

A method and device for predicting an air quality index. The method comprises the following steps: according to a geographical location to be predicted and a current time, acquiring a corresponding prediction model (S110); acquiring meteorological data of the geographical location to be predicted (S120); and according to the meteorological data, performing operation processing based on the prediction model, and determining an air quality index of the geographical location to be predicted at a prediction time (S130). The prediction of a future air quality index can be finished by means of only a relatively small amount of meteorological information which can be easily acquired, so that the implementation difficulty and an entry threshold of AQI prediction are reduced. In addition, since in the method, corresponding prediction models are established within all the different prediction time intervals with regard to each geographical location, climate characteristics of the local area can be better reflected and a powerful guarantee is provided for determining an AQI with relatively high accuracy.

Description

预测空气质量指数的方法与装置Method and device for predicting air quality index

技术领域Technical field

本发明涉及工业自动化控制技术领域，具体而言，本发明涉及一种预测空气质量指数的方法与装置。The present invention relates to the field of industrial automation control technology, and in particular, to a method and apparatus for predicting an air quality index.

背景技术Background technique

AQI(Air Quality Index，空气质量指数)是定量描述空气质量状况的无量纲指数。针对单项污染物还规定了空气质量分指数。参与空气质量评价的主要污染物为细颗粒物、可吸入颗粒物、二氧化硫、二氧化氮、臭氧、一氧化碳等六项。AQI (Air Quality Index) is a dimensionless index that quantitatively describes the state of air quality. The air quality sub-index is also specified for individual pollutants. The main pollutants involved in air quality assessment are fine particles, respirable particulate matter, sulfur dioxide, nitrogen dioxide, ozone, and carbon monoxide.

现有技术中，对AQI的主要预测方式为基于SPRINTARS大气环流模型的AQI预测方式。SPRINTARS(Spectral Radiation-Transport Model for Aerosol Species)是以全球规模模拟大气悬浮颗粒物，即大气气溶胶，对气候***造成的影响以及大气污染状况而开发的数值模型。SPRINTARS以东京大学大气海洋研究所(气候***研究系)、日本国立环境研究所及日本海洋研究开发机构(地球环境变动领域)开发的海气耦合模型MIROC为基础，对存在于对流层中的自然形成及人为形成的主要大气气溶胶，包括黑色碳、有机物、硫酸盐、土壤颗粒及海盐颗粒，进行研究，这些颗粒物也被分类为PM10及PM2.5。SPRINTARS可计算出大气气溶胶的移动过程，包括产生、对流、扩散、湿沉降、干沉降、重力沉降的过程；还可计算出大气气溶胶产生的直接作用，如大气气溶胶对阳光及红外线的散射和吸收，和间接作用，如大气气溶胶作为云凝结核及冰核的功能。In the prior art, the main prediction mode for AQI is the AQI prediction method based on the SPRINTARS atmospheric circulation model. SPRINTARS (Spectral Radiation-Transport Model for Aerosol Species) is a numerical model developed on a global scale to simulate atmospheric aerosols, ie atmospheric aerosols, effects on the climate system and atmospheric pollution. SPRINTARS is based on the sea-air coupled model MIROC developed by the Institute of Atmospheric and Oceanic Research (the Department of Climate Systems) of the University of Tokyo, the National Institute of Environmental Studies of Japan, and the Japan Ocean Research and Development Agency (the field of global environmental change), and the natural formation in the troposphere. And artificially formed major atmospheric aerosols, including black carbon, organic matter, sulfates, soil particles and sea salt particles, were also classified as PM10 and PM2.5. SPRINTARS can calculate the movement process of atmospheric aerosols, including production, convection, diffusion, wet deposition, dry deposition, gravity sedimentation; it can also calculate the direct effects of atmospheric aerosols, such as atmospheric aerosols for sunlight and infrared Scattering and absorption, and indirect effects, such as atmospheric aerosols, function as cloud condensation nuclei and ice cores.

SPRINTARS构造了一个相对完善、复杂的大气环流模型，以此来计算全球污染物颗粒的扩散趋势。该方式对季节分明的平原地区有很好的预测结果，但是对于地形复杂的盆地、丘陵、高原、山地等地区，可能会由于气候复杂，水汽兴盛，从而导致难以给出准确的预测。例如，SPRINTARS无法很好地区分中国西南部地区的雾霾与水雾，如对于少风多雾的中国重庆地区经常给出持续严重污染的预测，但实际该城市空气质量一直处于优良状态，使得基于SPRINTARS的AQI预报在中国西南部地区没有参考意义。SPRINTARS constructed a relatively complete and complex atmospheric circulation model to calculate the diffusion trend of global pollutant particles. This method has good prediction results for seasonally distinct plains, but for basins with complex terrain, hills, plateaus, mountains, etc., due to the complex climate and the prosperous water vapor, it is difficult to give accurate predictions. For example, SPRINTARS does not distinguish well between smog and water fog in southwestern China. For example, Chongqing, China, which has less wind and fog, often gives forecasts of persistent and serious pollution, but the city’s air quality has always been in excellent condition. The AQI forecast based on SPRINTARS has no reference significance in southwestern China.

因此，现有技术中，基于SPRINTARS大气环流模型的AQI预测方式存在以下问题：Therefore, in the prior art, the AQI prediction method based on the SPRINTARS atmospheric circulation model has the following problems:

1)SPRINTARS大气环流模型主要考虑大气环流整体因素，从大气环流的维度上分析污染物的扩散形式，而对于某个具体城市的具体气候情况难于详细区分。由于同一个城市的具体气候情况，会因不同季节、不同时间段、甚至人为因素而有所改变，例如，某地区新建了化工厂前后，污染物的排放和积累肯定会有所变化，因此，基于SPRINTARS大气环流模型的AQI预测方式对于具体城市的具体预测准确性较低。1) The SPRINTARS atmospheric circulation model mainly considers the overall factors of atmospheric circulation, and analyzes the diffusion form of pollutants from the dimension of atmospheric circulation, but it is difficult to distinguish the specific climate conditions of a specific city. Due to the specific climate of the same city, it will change due to different seasons, different time periods and even human factors. For example, before and after a new chemical plant is built in a certain area, the emission and accumulation of pollutants will definitely change. Therefore, the AQI prediction method based on the SPRINTARS atmospheric circulation model is less accurate for specific cities.

2)SPRINTARS大气环流模型的数据采集量非常巨大，至少需要收集大量的污染源具体信息及卫星气象信息，同时，SPRINTARS大气环流模型的计算量也是非常庞大，需要高性能的硬件设备来支持，巨大的数据采集量及庞大的计算量对于普通的单位和个人存在很高的技术进入门槛。2) The data collection of SPRINTARS atmospheric circulation model is very large. At least a large amount of pollution source specific information and satellite meteorological information need to be collected. At the same time, the calculation amount of SPRINTARS atmospheric circulation model is also very large, which requires high-performance hardware equipment to support, huge The amount of data collected and the large amount of calculations have high technical entry barriers for ordinary units and individuals.

发明内容Summary of the invention

为克服上述技术问题或者至少部分地解决上述技术问题，特提出以下技术方案：In order to overcome the above technical problems or at least partially solve the above technical problems, the following technical solutions are proposed:

本发明的实施例提出了一种预测空气质量指数的方法，包括：Embodiments of the present invention provide a method of predicting an air quality index, including:

根据待预测地理位置及当前时间，获取相应的预测模型；Obtain a corresponding prediction model according to the geographical location to be predicted and the current time;

获取所述待预测地理位置的气象数据；Obtaining meteorological data of the geographical location to be predicted;

根据所述气象数据信息，基于所述预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。Based on the weather data information, an arithmetic process is performed based on the prediction model to determine an air quality index of a to-be-predicted geographic location at a time to be predicted.

本发明的实施例还提出了一种预测空气质量指数的装置，包括：Embodiments of the present invention also provide an apparatus for predicting an air quality index, comprising:

模型获取模块，用于根据待预测地理位置及当前时间，获取相应的预测模型；a model obtaining module, configured to acquire a corresponding prediction model according to the geographic location to be predicted and the current time;

气象数据获取模块，用于获取所述待预测地理位置的气象数据；a meteorological data acquisition module, configured to acquire weather data of the geographic location to be predicted;

指数确定模块，用于根据所述气象数据信息，基于所述预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。And an index determining module, configured to perform, according to the meteorological data information, an operation process based on the prediction model to determine an air quality index of a to-be-predicted geographic location at a time to be predicted.

本发明的实施例还提出了一种程序，包括可读代码，当所述可读代码在计算设备上运行时，导致所述计算设备执行本发明实施例所述的预测空气质量指数的方法。Embodiments of the present invention also provide a program comprising readable code that, when executed on a computing device, causes the computing device to perform the method of predicting an air quality index as described in embodiments of the present invention.

本发明的实施例还提出了一种可读介质，其中存储了上述程序。Embodiments of the present invention also propose a readable medium in which the above program is stored.

本发明的实施例中，可针对每一地理位置在当前时间下均建立相应的预测模型；在执行预测时，首先，获取待预测地理位置的气象数据，该气象数据可从气象部门发布的天气预报等公共信息中获取；随后，根据气象数据，基于相应的预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。本发明相对于现有技术的优势在于：首先，本发明仅需较少的且易于获取的气象信息即可完成未来空气质量指数的预测，由于本发明的方案不需要获取现有技术中大气环流模型所需的复杂大气环流参数，而仅需易于获得的基本气象数据，如历史每日温度、湿度、气压、风力、风向、AQI数值等，即可完成预测，因此任何单位和个人都可以利用气象部门发布的公共气象信息基于本发明所述的技术方案进行AQI预测，同时，基于预测模型，如基于人工神经网络的预测模型的计算量较小，无需高性能硬件设备的高成本投入，降低了预测的实现难度和进入门槛。另外，由于本发明针对每一地理位置在不同的预测时间区间下均建立相应的预测模型，从而可以更好地反映本地区的气候特点，相对于现有技术的大气环流模型有着更好的区域适应性，为确定精确度较高的AQI提供有力保证。In an embodiment of the present invention, a corresponding prediction model may be established for each geographic location at a current time; when performing prediction, first, acquiring meteorological data of a geographic location to be predicted, the weather data may be released from weather by the meteorological department. Obtaining in public information such as forecasting; subsequently, based on the meteorological data, performing arithmetic processing based on the corresponding prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted. The advantages of the present invention over the prior art are: First, the present invention requires only less and easily accessible meteorological information to predict future air quality indices, since the solution of the present invention does not require acquisition of atmospheric circulation in the prior art. The complex atmospheric circulation parameters required by the model, and only basic weather data that are easy to obtain, such as historical daily temperature, humidity, air pressure, wind, wind direction, AQI values, etc., can be used to predict, so that any unit and individual can use The public meteorological information issued by the meteorological department performs AQI prediction based on the technical solution described in the present invention, and at the same time, based on the prediction model, such as the artificial neural network-based prediction model, the calculation amount is small, and high-cost input of high-performance hardware equipment is not required, and the reduction is required. The difficulty of implementation and the threshold for entry. In addition, since the present invention establishes a corresponding prediction model for each geographic location in different prediction time intervals, thereby better reflecting the climate characteristics of the region, there is a better region than the prior art atmospheric circulation model. Adaptability, providing a strong guarantee for determining the accuracy of AQI certificate.

本发明附加的方面和优点将在下面的描述中部分给出，这些将从下面的描述中变得明显，或通过本发明的实践了解到。The additional aspects and advantages of the invention will be set forth in part in the description which follows.

附图说明DRAWINGS

本发明上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解，其中：The above and/or additional aspects and advantages of the present invention will become apparent and readily understood from

图1为本发明中的预测空气质量指数的方法的一个实施例的流程示意图；1 is a schematic flow chart of one embodiment of a method for predicting an air quality index in the present invention;

图2为本发明中预测空气质量指数的方法一个优选实施例的流程示意图；2 is a schematic flow chart of a preferred embodiment of a method for predicting an air quality index in the present invention;

图3为本发明中的预测空气质量指数一个具体实施例的流程示意图；3 is a schematic flow chart of a specific embodiment of a predicted air quality index in the present invention;

图4为本发明中预测空气质量指数的装置一个实施例的结构示意图；4 is a schematic structural view of an embodiment of a device for predicting an air quality index according to the present invention;

图5为本发明中预测空气质量指数的装置一个优选实施例的结构示意图；Figure 5 is a schematic view showing the structure of a preferred embodiment of the apparatus for predicting the air quality index in the present invention;

图6示出了用于执行根据本发明的预测空气质量指数的方法的计算设备的框图；6 shows a block diagram of a computing device for performing a method of predicting an air quality index in accordance with the present invention;

图7示出了用于保持或者携带实现根据本发明的预测空气质量指数的方法的程序代码的存储单元。Figure 7 shows a storage unit for program code for maintaining or carrying a method of implementing a predicted air quality index in accordance with the present invention.

具体实施方式detailed description

下面详细描述本发明的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，仅用于解释本发明，而不能解释为对本发明的限制。The embodiments of the present invention are described in detail below, and the examples of the embodiments are illustrated in the drawings, wherein the same or similar reference numerals are used to refer to the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the drawings are intended to be illustrative of the invention and are not to be construed as limiting.

本技术领域技术人员可以理解，除非特意声明，这里使用的单数形式“一”、“一个”、“所述”和“该”也可包括复数形式。应该进一步理解的是，本发明的说明书中使用的措辞“包括”是指存在所述特征、整数、步骤、操作、元件和/或组件，但是并不排除存在或添加一个或多个其他特征、整数、步骤、操作、元件、组件和/或它们的组。应该理解，当我们称元件被“连接”或“耦接”到另一元件时，它可以直接连接或耦接到其他元件，或者也可以存在中间元件。此外，这里使用的“连接”或“耦接”可以包括无线连接或无线耦接。这里使用的措辞“和/或”包括一个或更多个相关联的列出项的全部或任一单元和全部组合。The singular forms "a", "an", "the" It is to be understood that the phrase "comprise" or "an" Integers, steps, operations, components, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element. Further, "connected" or "coupled" as used herein may include either a wireless connection or a wireless coupling. The phrase "and/or" used herein includes all or any one and all combinations of one or more of the associated listed.

本技术领域技术人员可以理解，除非另外定义，这里使用的所有术语(包括技术术语和科学术语)，具有与本发明所属领域中的普通技术人员的一般理解相同的意义。还应该理解的是，诸如通用字典中定义的那些术语，应该被理解为具有与现有技术的上下文中的意义一致的意义，并且除非像这里一样被特定定义，否则不会用理想化或过于正式的含义来解释。Those skilled in the art will appreciate that all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the invention belongs, unless otherwise defined. It should also be understood that terms such as those defined in a general dictionary should be understood to have meaning consistent with the meaning in the context of the prior art, and will not be idealized or excessive unless specifically defined as here. The formal meaning is explained.

在上下文中所称“计算机设备”，也称为“电脑”，是指可以通过运行预定程序或指令来执行数值计算和/或逻辑计算等预定处理过程的智能电子设备，其可以包括处理器与存储器，由处理器执行在存储器中预存的存续指令来执行预定处理过程，或是由ASIC、FPGA、DSP等硬件执行预定处理过程，或是由上述二者组合来实现。计算机设备包括但不限于服务器、个人电脑、笔记本电脑、平板电脑、智能手机等。By "computer device", also referred to as "computer" in the context, is meant an intelligent electronic device that can perform predetermined processing, such as numerical calculations and/or logical calculations, by running predetermined programs or instructions, which may include The processor and the memory are executed by the processor to execute a predetermined process pre-stored in the memory, or perform a predetermined process by hardware such as an ASIC, an FPGA, a DSP, or the like, or a combination of the two. Computer devices include, but are not limited to, servers, personal computers, notebook computers, tablets, smart phones, and the like.

所述计算机设备包括用户设备与网络设备。其中，所述用户设备包括但不限于电脑、智能手机、PDA等；所述网络设备包括但不限于单个网络服务器、多个网络服务器组成的服务器组或基于云计算(Cloud Computing)的由大量计算机或网络服务器构成的云，其中，云计算是分布式计算的一种，由一群松散耦合的计算机集组成的一个超级虚拟计算机。其中，所述计算机设备可单独运行来实现本发明，也可接入网络并通过与网络中的其他计算机设备的交互操作来实现本发明。其中，所述计算机设备所处的网络包括但不限于互联网、广域网、城域网、局域网、VPN网络等。The computer device includes a user device and a network device. The user equipment includes, but is not limited to, a computer, a smart phone, a PDA, etc.; the network device includes but is not limited to a single network server, a server group composed of multiple network servers, or a cloud computing based computer Or a cloud composed of a network server, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computers. Wherein, the computer device can be operated separately to implement the present invention, and can also access the network and implement the present invention by interacting with other computer devices in the network. The network in which the computer device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a VPN network, and the like.

所述人工神经网络是一种应用类似于大脑神经突触联接的结构进行信息处理的数学模型，在工程与学术界也常直接简称为神经网络或类神经网络。人工神经网络是一种运算模型，由大量的节点，或称神经元，之间相互连接构成。每个节点代表一种特定的输出函数，称为激励函数(activation function)。每两个节点间的连接都代表一个对于通过该连接信号的加权值，称之为权重，相当于人工神经网络的记忆。人工神经网络的输出则基于人工神经网络的连接方式、节点间权重和激励函数的不同而不同。而人工神经网络自身通常都是对自然界某种算法或者函数的逼近，也可能是对一种逻辑策略的表达。The artificial neural network is a mathematical model for applying information processing similar to the structure of the brain synaptic connection, and is often referred to simply as a neural network or a neural network in engineering and academia. An artificial neural network is an operational model consisting of a large number of nodes, or neurons, connected to each other. Each node represents a specific output function called an activation function. The connection between every two nodes represents a weighting value for the signal passing through the connection, called weight, which is equivalent to the memory of the artificial neural network. The output of the artificial neural network is different based on the connection mode of the artificial neural network, the weight between nodes, and the excitation function. The artificial neural network itself is usually an approximation of an algorithm or function in nature, or it may be an expression of a logic strategy.

图1为本发明中的预测空气质量指数的方法的一个实施例的流程示意图。1 is a schematic flow chart of one embodiment of a method for predicting an air quality index in the present invention.

本方法由计算机设备执行；在步骤S110中，计算机设备根据待预测地理位置及当前时间，获取相应的预测模型；在步骤S120中，获取待预测地理位置的气象数据；在步骤S130中，根据气象数据，并基于预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。The method is executed by the computer device; in step S110, the computer device acquires a corresponding prediction model according to the geographic location to be predicted and the current time; in step S120, the meteorological data of the geographic location to be predicted is acquired; in step S130, according to the meteorological data The data is processed based on the prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted.

本发明的实施例中，可针对每一地理位置在当前时间下均建立相应的预测模型；在执行预测时，首先，获取待预测地理位置的气象数据，该气象数据可从气象部门发布的天气预报等公共信息中获取；随后，根据气象数据，基于相应的预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。本发明相对于现有技术的优势在于：首先，本发明仅需较少的且易于获取的气象信息即可完成未来空气质量指数的预测，由于本发明的方案不需要获取现有技术中大气环流模型所需的复杂大气环流参数，而仅需易于获得的基本气象数据，如历史每日温度、湿度、气压、风力、风向、AQI数值等，即可完成预测，因此任何单位和个人都可以利用气象部门发布的公共气象信息基于本发明所述的技术方案进行AQI预测，同时，基于预测模型，如基于人工神经网络的预测模型的计算量较小，无需高性能硬件设备的高成本投入，降低了预测的实现难度和进入门槛。另外，由于本发明针对每一地理位置在不同的预测时间区间下均建立相应的预测模型，从而可以更好地反映本地区的气候特点，相对于现有技术的大气环流模型有着更好的区域适应性，为确定精确度较高的AQI提供有力保证。In an embodiment of the present invention, a corresponding prediction model may be established for each geographic location at a current time; when performing prediction, first, acquiring meteorological data of a geographic location to be predicted, the weather data may be released from weather by the meteorological department. Obtaining in public information such as forecasting; subsequently, based on the meteorological data, performing arithmetic processing based on the corresponding prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted. The advantages of the present invention over the prior art are: First, the present invention requires only less and easily accessible meteorological information to predict future air quality indices, since the solution of the present invention does not require acquisition of atmospheric circulation in the prior art. The complex atmospheric circulation parameters required by the model, and only basic weather data that are easy to obtain, such as historical daily temperature, humidity, air pressure, wind, wind direction, AQI values, etc., can be used to predict, so that any unit and individual can use The public meteorological information issued by the meteorological department performs AQI prediction based on the technical solution described in the present invention, and at the same time, based on the prediction model, such as the artificial neural network-based prediction model, the calculation amount is small, and high-cost input of high-performance hardware equipment is not required, and the reduction is required. The difficulty of implementation and the threshold for entry. In addition, since the present invention establishes a corresponding prediction model for each geographic location in different prediction time intervals, the climate characteristics of the region can be better reflected, as opposed to the present The technical atmospheric circulation model has better regional adaptability and provides a strong guarantee for determining the accuracy of AQI.

本方案中，预测模型可为“基于人工神经网络”的预测模型，以下将以基于人工神经网络的预测模型作为一个实现方式进行具体实施例的撰写。In this solution, the prediction model may be an "artificial neural network-based" prediction model. The following is a description of a specific embodiment using an artificial neural network-based prediction model as an implementation.

在步骤S110中，计算机设备根据待预测地理位置及当前时间，获取相应的预测模型。In step S110, the computer device acquires a corresponding prediction model according to the geographic location to be predicted and the current time.

具体地，建立预测模型的过程可为实时建立及预先建立。Specifically, the process of establishing a predictive model can be established and pre-established in real time.

当建立基于人工神经网络的预测模型的过程为实时建立时，步骤S110(参照图1)包括步骤S111(图中未示出)、步骤S112(图中未示出)、步骤S113(图中未示出)和步骤S114(图中未示出)；在步骤S111中，确定当前时间所在的预测时间区间；在步骤S112中确定与待预测地理位置及所述预测时间区间相匹配的气象指标；在步骤S113中，根据待预测地理位置及所述预测时间区间，获取气象指标下的历史气象样本数据；在步骤S114中，根据历史气象样本数据进行机器学习，确定与待预测地理位置及当前时间相对应的预测模型；其中，步骤S111、步骤S112、步骤S113与步骤S114中的实时建立预测模型的方式与预先建立预测模型的方式相似，在此参考以下述实施例的具体实施方式。When the process of establishing the artificial neural network-based prediction model is established in real time, step S110 (refer to FIG. 1) includes step S111 (not shown), step S112 (not shown), step S113 (not shown) And step S114 (not shown); in step S111, determining a predicted time interval in which the current time is located; determining a weather indicator matching the predicted geographical position and the predicted time interval in step S112; In step S113, historical meteorological sample data under the meteorological indicator is acquired according to the geographical position to be predicted and the predicted time interval; in step S114, machine learning is performed according to the historical weather sample data, and the geographical position and current time to be predicted are determined. Corresponding prediction models; wherein the manner of establishing the prediction model in real time in steps S111, S112, S113, and S114 is similar to the manner in which the prediction model is pre-established, and reference is made to the specific embodiments of the following embodiments.

优选地(参照图1)，该方法还包括：判断待预测地理位置在预测时间区间内是否存在样本过滤规则；若存在，在步骤S113中，根据待预测地理位置及预测时间区间，基于样本过滤规则，获取气象指标下的历史气象样本数据。Preferably (refer to FIG. 1), the method further comprises: determining whether there is a sample filtering rule in the predicted time interval in the predicted time interval; if present, filtering the sample based on the to-be-predicted geographic location and the predicted time interval in step S113 Rules to obtain historical meteorological sample data under meteorological indicators.

例如，北京地区在2014年11月上旬实施大面积停工及车辆限行等政策，导致污染物的排放量大幅度下降，则预置在选取北京地区11月期间的历史气象样本数据时，需将11月1日-10日的气象样本数据过滤。For example, in Beijing in the first ten days of November 2014, policies such as large-scale shutdowns and vehicle restrictions were implemented, resulting in a significant drop in pollutant emissions. The historical meteorological data for the period of November in Beijing should be preset. The weather sample data of January 1st - 10th is filtered.

当建立基于人工神经网络的预测模型的过程为预先建立时，步骤S110(参照图1)包括步骤S115(图中未示出)和步骤S116(图中未示出)；在步骤S115中，确定当前时间所在的预测时间区间；在步骤S116中，根据待预测地理位置及预测时间区间，在预测模型库中进行匹配查询，获取与待预测地理位置及预测时间区间相对应的预测模型。When the process of establishing the artificial neural network-based prediction model is pre-established, step S110 (refer to FIG. 1) includes step S115 (not shown) and step S116 (not shown); in step S115, determining The prediction time interval in which the current time is located; in step S116, the matching query is performed in the prediction model library according to the geographic location to be predicted and the prediction time interval, and the prediction model corresponding to the geographical position to be predicted and the prediction time interval is obtained.

优选地，预先建立各个地理位置在不同预测时间区间内相应的预测模型，将预测模型存储于预测模型库，并保存地理位置、预测时间区间及预测模型之间的对应关系，如将三者之间的对应关系保存于模型对应列表中，以供查询。Preferably, the corresponding prediction models of the respective geographical locations in different prediction time intervals are pre-established, the prediction models are stored in the prediction model library, and the correspondence between the geographical location, the prediction time interval and the prediction model is saved, for example, the three The correspondence between the two is stored in the model correspondence list for query.

如图2所示，在一优选实施例中，根据待预测地理位置及预测时间区间，在预测模型库中进行匹配查询，获取与待预测地理位置及当前时间相对应的预测模型之前，还包括步骤S240、步骤S250、步骤S260及步骤S270；在步骤S240中，确定与待预测地理位置及预测时间区间相匹配的气象指标；在步骤S250中，根据待预测地理位置及预测时间区间，获取气象指标下的历史气象样本数据；在步骤S260中，根据历史气象样本数据进行机器学习，确定与待预测地理位置及预测时间区间相对应的预测模型；在步骤S270中，将与待预测地理位置及预测时间区间相对应的预测模型保存至预测模型库。As shown in FIG. 2, in a preferred embodiment, according to the geographic location to be predicted and the predicted time interval, a matching query is performed in the prediction model library to obtain a prediction model corresponding to the geographic location to be predicted and the current time, and includes Step S240, step S250, step S260, and step S270; in step S240, determining a weather indicator that matches the predicted geographic location and the predicted time interval; and in step S250, acquiring the weather according to the predicted geographic location and the predicted time interval Historical weather sample data under the indicator; in step S260, the root Performing machine learning according to historical weather sample data, determining a prediction model corresponding to the predicted geographic location and the predicted time interval; and in step S270, saving the prediction model corresponding to the predicted geographic location and the predicted time interval to the prediction model library .

优选地，当预测模型库已存在与待预测地理位置及预测时间区间相对应的预测模型时，用最新的与待预测地理位置及预测时间区间相对应的预测模型替换先前已存在的与待预测地理位置及预测时间区间相对应的预测模型。Preferably, when the prediction model library already has a prediction model corresponding to the to-be-predicted geographic location and the prediction time interval, the prediction model corresponding to the to-be-predicted geographic location and the prediction time interval is replaced with the previously existing and to-be-predicted The prediction model corresponding to the geographic location and the predicted time interval.

在步骤S240中，确定与待预测地理位置及预测时间区间相匹配的气象指标；In step S240, determining a weather indicator that matches the geographic location to be predicted and the predicted time interval;

其中，气象指标包括但不限于：温度指标，可包括日最高温度、日最低温度；湿度指标，可包括日平均湿度；风力指标，可包括日最大风力、日主导风向的最大风力、日主导风向的平均风力；风向指标，可包括日最大风向；气压指标，可包括日平均气压；降雨量指标，可包括日平均降雨量、日最大降雨量；露点指标，可包括日平均露点；空气质量指数指标，即AQI。The meteorological indicators include, but are not limited to, temperature indicators, which may include daily maximum temperature and daily minimum temperature; humidity indicators, which may include daily average humidity; wind indicators, which may include daily maximum wind force, maximum wind force of daily dominant wind direction, and daily dominant wind direction. Average wind; wind direction indicators, which may include daily maximum wind direction; air pressure indicators, which may include daily average air pressure; rainfall indicators, which may include daily average rainfall, daily maximum rainfall; dew point indicators, which may include daily average dew point; air quality index Indicator, AQI.

例如，待预测地理位置为北京地区，预测时间区间为“1-3月”，在下表1中进行查询匹配，确定与“北京地区”及预测时间区间“1-3月”相匹配的气象指标包括日最高温度、日最低温度、日平均湿度、日主导风向的最大风力以及AQI。For example, if the predicted geographical position is Beijing, the forecast time interval is “1-3 months”, and the query matching is performed in Table 1 below to determine the meteorological indicators that match the “Beijing area” and the forecast time interval “1-3 months”. Including daily maximum temperature, daily minimum temperature, daily average humidity, maximum wind force for dominant wind direction, and AQI.

表1：Table 1:

在步骤S250中，根据待预测地理位置及预测时间区间，获取气象指标下的历史气象样本数据。In step S250, the history under the meteorological indicator is obtained according to the geographical position to be predicted and the predicted time interval. Meteorological sample data.

例如，接上例，待预测地理位置为北京地区，预测时间区间为“1-3月”，且相匹配的气象指标包括日最高温度、日最低温度、日平均湿度、日主导风向的最大风力以及AQI，故获取北京地区2015年1-3月及2012-2014年1-3月的历史的每日最高温度、每日最低温度、每日平均湿度、每日主导风向的最大风力以及每日AQI作为历史气象样本数据；其中，获取历史气象样本数据时需考虑季节等因素，获取历史气象样本数据的标准可包括：近期的气象样本数据，如近3个月的气象样本数据，以及历史同期的气象样本数据，如近3年的历史同期的气象样本数据。For example, in the above example, the predicted geographical position is Beijing, the forecast time interval is “1-3 months”, and the matching meteorological indicators include the daily maximum temperature, the daily minimum temperature, the daily average humidity, and the maximum wind force of the dominant wind direction. As well as AQI, the daily maximum temperature, daily minimum temperature, daily average humidity, daily maximum wind force and daily maximum wind force of the Beijing area from January to March of 2015 and January to January of 2012-2014 are obtained. AQI is used as historical meteorological sample data; among them, factors such as seasonality should be taken into account when obtaining historical meteorological sample data. The criteria for obtaining historical meteorological sample data may include: recent meteorological sample data, such as meteorological sample data of the past 3 months, and historical period Meteorological sample data, such as meteorological sample data for the same period of the past 3 years.

优选地(参照图2)，该方法还包括步骤S280(图中未示出)；在步骤S280中，判断待预测地理位置在预测时间区间内是否存在样本过滤规则；若存在，在步骤S250中，根据待预测地理位置及预测时间区间，基于样本过滤规则，获取气象指标下的历史气象样本数据。Preferably (refer to FIG. 2), the method further includes step S280 (not shown); in step S280, determining whether the predicted geographic location has a sample filtering rule within the predicted time interval; if present, in step S250 According to the geographical location to be predicted and the predicted time interval, the historical meteorological sample data under the meteorological indicator is obtained based on the sample filtering rule.

在步骤S260中，根据历史气象样本数据进行机器学习，确定与待预测地理位置及预测时间区间相对应的预测模型。In step S260, machine learning is performed based on the historical weather sample data, and a prediction model corresponding to the geographical position to be predicted and the predicted time interval is determined.

具体地，步骤S260(参照图2)包括步骤S261(图中未示出)和步骤S262(图中未示出)；在步骤S261中，根据历史气象样本数据，基于人工神经网络进行机器学习，确定与待预测地理位置及预测时间区间相对应的人工神经网络的节点间权重；在步骤S262中，根据人工神经网络的节点间权重，建立相对应的预测模型。Specifically, step S260 (refer to FIG. 2) includes step S261 (not shown) and step S262 (not shown); in step S261, based on the historical weather sample data, machine learning is performed based on the artificial neural network. Determining the inter-node weights of the artificial neural network corresponding to the predicted geographical position and the predicted time interval; and in step S262, establishing a corresponding prediction model according to the inter-node weights of the artificial neural network.

具体地，将历史气象样本数据作为人工神经网络的输入数据，基于人工神经网络进行机器学习，确定机器学习结束后人工神经网络的节点间权重；随后，根据人工神经网络的节点间权重，建立相对应的预测模型。Specifically, the historical meteorological sample data is used as the input data of the artificial neural network, and the machine learning is performed based on the artificial neural network to determine the inter-node weight of the artificial neural network after the end of the machine learning; then, the phase is established according to the weight among the nodes of the artificial neural network. Corresponding prediction model.

其中，步骤S261(图中未示出)包括步骤S2611(图中未示出)、步骤S2612(图中未示出)和步骤S2613(图中未示出)：在步骤S2611中，根据历史气象样本数据，基于人工神经网络进行机器学习，确定人工神经网络的空气质量指数学习结果；在步骤S2612中，计算空气质量指数学习结果与历史气象样本数据中的历史空气质量指数的误差值；在步骤S2613中，当误差值小于预定误差阈值时，提取人工神经网络的节点间权重。Wherein, step S261 (not shown) includes step S2611 (not shown), step S2612 (not shown), and step S2613 (not shown): in step S2611, according to historical weather Sample data, based on artificial neural network for machine learning, determining an air quality index learning result of the artificial neural network; in step S2612, calculating an error value of the air quality index learning result and the historical air quality index in the historical weather sample data; In S2613, when the error value is less than the predetermined error threshold, the inter-node weight of the artificial neural network is extracted.

具体地，将历史气象样本数据作为人工神经网络的输入数据，基于人工神经网络进行机器学习，获取人工神经网络的输出数据，即空气质量指数学习结果；接着，计算输出数据与输入数据中的空气质量指数的误差值，即空气质量指数学习结果与历史气象样本数据中的空气质量指数的误差值；当误差值小于预定误差阈值时，确定机器学习结束，提取人工神经网络的节点间权重。 Specifically, the historical meteorological sample data is used as the input data of the artificial neural network, and the machine learning is performed based on the artificial neural network to obtain the output data of the artificial neural network, that is, the air quality index learning result; then, the air in the output data and the input data is calculated. The error value of the quality index, that is, the error value of the air quality index learning result and the air quality index in the historical weather sample data; when the error value is less than the predetermined error threshold, it is determined that the machine learning ends, and the inter-node weight of the artificial neural network is extracted.

在步骤S270中，将与待预测地理位置及预测时间区间相对应的预测模型保存至预测模型库。In step S270, the prediction model corresponding to the geographic location to be predicted and the predicted time interval is saved to the prediction model library.

随着不断有新的气象样本数据产生，每隔预定更新时间间隔，可循环执行一次机器学习，针对每一地理位置提取新的人工神经网络的节点间权重，并存储在更新预测模型库中，具体可为针对每一地理位置提取重新机器学习后的人工神经网络的节点间权重，具体可为在预测模型库中删除原有的节点间权重的数据记录，写入最新的数据。With the continuous generation of new meteorological sample data, machine learning can be performed cyclically every predetermined update interval, and the inter-node weights of the new artificial neural network are extracted for each geographic location and stored in the updated prediction model library. Specifically, the inter-node weights of the artificial neural network after re-machine learning may be extracted for each geographic location, specifically, the data records of the original inter-node weights are deleted in the prediction model library, and the latest data is written.

在具体机器学***原地区受风力影响很大，可以使用风指标来选取样本数据。因此，优选实施例中，人工神经网络是根据不同地理位置在不同预测时间区间内的历史气象样本数据进行机器学习而得到的，即每一地理位置都有一套独一无二的人神经网络的节点间权重。每一地理位置都有一套适于本地区的人工神经网络，可以更好地反映本地区的气候特点，为确定精确度较高的AQI提供有力保证。In the process of specific machine learning, according to the actual situation of different geographical locations in different prediction time intervals, different meteorological indicators are selected to select historical meteorological sample data for machine learning. For example, if the wind direction in the mountainous area is uncertain and the wind is small, the wind index can be removed to prevent interference with machine learning. The plain area is greatly affected by the wind, and the wind index can be used to select the sample data. Therefore, in the preferred embodiment, the artificial neural network is obtained by machine learning based on historical meteorological sample data of different geographic locations in different prediction time intervals, that is, each geographic location has a unique set of internode weights of the human neural network. . Each geographic location has an artificial neural network suitable for the region, which can better reflect the climate characteristics of the region and provide a strong guarantee for determining the accuracy of the AQI.

参照图1，在步骤S120中，获取待预测地理位置的气象数据。Referring to FIG. 1, in step S120, weather data of a geographic location to be predicted is acquired.

具体地，步骤S120包括步骤S121(图中未示出)、步骤S122(图中未示出)和步骤S123(图中未示出)；在步骤S121中，确定当前时间所在的预测时间区间；在步骤S122中，确定与待预测地理位置及所述预测时间区间相匹配的气象指标；在步骤S123中，根据待预测地理位置，确定气象指标下的气象数据。Specifically, step S120 includes step S121 (not shown in the figure), step S122 (not shown in the figure), and step S123 (not shown); in step S121, determining a predicted time interval in which the current time is located; In step S122, a weather indicator matching the to-be-predicted geographical position and the predicted time interval is determined; in step S123, the meteorological data under the meteorological indicator is determined according to the geographical position to be predicted.

具体地，首先，根据上表1，确定当前时间对应的预测时间区间，接着，确定与待预测地理位置及预测时间区间相匹配的气象指标；随后，根据待预测地理位置，确定气象指标下的气象数据。Specifically, first, according to the above Table 1, determining a predicted time interval corresponding to the current time, and then determining a weather indicator that matches the to-be-predicted geographical position and the predicted time interval; and subsequently, determining the meteorological indicator according to the geographical position to be predicted Meteorological data.

当待预测时间为明日时，气象数据包括：今日的气象数据及明日的预测气象数据；或，前预定天数的真实气象数据，今日的气象数据及明日的预测气象数据；其中，今日的气象数据包括：今日的真实气象数据；如果今日的真实气象数据不完整，则可包括今日的真实气象数据与今日的预测气象数据。其中，预测气象数据包括但不限于：温度信息；湿度信息；风向信息；风力信息；气压信息；降雨量信息；露点信息。When the forecasting time is tomorrow, the meteorological data includes: today's meteorological data and forecasted meteorological data for tomorrow; or real weather data for the previous scheduled days, today's meteorological data and tomorrow's forecasted meteorological data; among them, today's meteorological data Including: true weather data today; if today's real weather data is incomplete, it can include today's real weather data and today's forecast weather data. Among them, the predicted meteorological data includes but is not limited to: temperature information; humidity information; wind direction information; wind information; air pressure information; rainfall information; dew point information.

例如，待预测时间为明日；则气象数据可包括今日真实的空气质量指数、温度信息、湿度信息、风向信息、风力信息、气压信息、降雨量信息、露点信息等，以及明日预测的温度信息、湿度信息、风向信息、风力信息、气压信息、降雨量信息、露点信息等。 For example, the time to be predicted is tomorrow; the weather data may include today's real air quality index, temperature information, humidity information, wind direction information, wind information, air pressure information, rainfall information, dew point information, etc., as well as temperature information predicted tomorrow, Humidity information, wind direction information, wind information, air pressure information, rainfall information, dew point information, etc.

在一示例中，待预测地理位置为“北京地区”，今日为2015年1月14日，待预测时间为明日2015年1月15日，则可通过气象局发布的天气预报获取2015年1月14日当日的真实的最高温度、最低温度、相对湿度、最大风速、最大风速风向、AQI等真实气象数据，并可获取2015年1月15日预测的最高温度、最低温度、相对湿度、最大风速、最大风速风向等预测气象数据。In an example, the geographic location to be predicted is “Beijing Area”, today is January 14, 2015, and the forecast time is tomorrow January 15, 2015. The weather forecast published by the Meteorological Bureau can be obtained in January 2015. The true maximum temperature, minimum temperature, relative humidity, maximum wind speed, maximum wind speed and direction, AQI and other real weather data on the 14th, and can obtain the highest temperature, minimum temperature, relative humidity and maximum wind speed predicted on January 15, 2015. Forecasting meteorological data such as maximum wind speed and direction.

在步骤S130中，根据气象数据，基于预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。In step S130, based on the meteorological data, an arithmetic processing is performed based on the prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted.

例如，如表2所示，2015年1月14日是北京地区来自实际观测的当日真实气象数据，2015年1月15日及以后日期是来自天气预报的预测气象数据。For example, as shown in Table 2, January 14, 2015 is the actual weather data from the actual observations in Beijing, and the date of January 15, 2015 and later is the predicted meteorological data from the weather forecast.

表2Table 2

时间time	AQIAQI	最高温度Maximum temperature	最低温度lowest temperature	相对湿度Relative humidity	最大风速Maximum wind speed	最大风速风向Maximum wind speed and direction
2015/1/142015/1/14	267267	33	-6-6	6969	88	340340
2015/1/152015/1/15		66	-5-5	7979	1616	155155
2015/1/162015/1/16		44	-7-7	4141	4040	329329
2015/1/172015/1/17		66	-4-4	3333	1616	223223
2015/1/182015/1/18		99	-5-5	3636	24twenty four	283283
2015/1/192015/1/19		88	-6-6	4343	1616	113113
2015/1/202015/1/20		99	-4-4	5151	1616	330330
2015/1/212015/1/21		77	-6-6	3737	1616	258258
2015/1/222015/1/22		88	-4-4	3939	1616	268268
2015/1/232015/1/23		77	-5-5	3131	1616	290290

待预测地理位置为“北京地区”，今日为2015年1月14日，气象数据如上表2所示，根据2015年1月14日今日的真实气象数据及2015年1月15日的明日的预测气象数据，基于与该地理位置及当前时间相匹配的预测模型进行运算处理，确定北京地区在2015年1月15日明日的空气质量指数。The predicted geographical location is “Beijing Area”, today is January 14, 2015, and the meteorological data is as shown in Table 2 above, based on today's real weather data on January 14, 2015 and tomorrow's forecast on January 15, 2015. The meteorological data is calculated based on the prediction model matching the geographical location and the current time to determine the air quality index of the Beijing area on January 15, 2015 tomorrow.

如图3所示，在一具体应用场景中，在预测模型的预建阶段，待预测地理位置为城市A，预先执行机器学习来获取城市A的基于人工神经网络的预测模型；首先查询获取城市A的历史气候样本数据，判断针对该城市A是否存在样本过滤规则，若没有，则根据具体的预测时间区间获取城市A默认的历史气候样本数据，并作为人工神经网络的输入数据；若有，则根据具体的预测时间区间及样本过滤规则，获取城市A的过滤后的历史气候样本数据，并作为人工神经网络的输入数据；随后，该历史气候样本数据在人工神经网络进行机器学习训练，获取人工神经网络的输出数据，输出数据为空气质量指数学习结果；接着，比较人工神经网络的输入数据中的历史空气质量指数与输出数据的空气质量指数学习结果的误差值，若误差值大于预定阈值时，继续执行机器学习；若误差值小于预定阈值时，机器学习完毕，输出并存储城市A在该预测时间区间的节点间权重，从而建立城市A在该预测时间区间内的预测模型。在预测阶段，首先读取城市A的气象数据，根据当前时间加载相匹配的城市A的预测模型，并将城市A的气象数据作为人工神经网络的输入数据，通过人工神经网络进行运算处理，获得人工神经网络的输出数据，即为预测AQI值，随后利用最新获取的AQI，循环执行该运算步骤。例如，如表2所示，第一轮计算过程，根据2015年1月14日今日真实的最高温度、最低温度、相对湿度、最大风速、最大风速风向、AQI以及2015年1月15日明日的预测的最高温度、最低温度、相对湿度、最大风速、最大风速风向，通过基于人工神经网络的预测模型来运算确定2015年1月15明日日的预测AQI；第二轮计算过程，利用2015年1月15和2015年1月16日的预测的最高温度、最低温度、相对湿度、最大风速、最大风速风向，及预测得到的2015年1月15日的AQI，通过运算确定2015年1月16日的预测AQI；以此类推进行运算，可确定15～22日的AQI预测结果。As shown in FIG. 3, in a specific application scenario, in the pre-construction phase of the prediction model, the geographic location to be predicted is city A, and machine learning is performed in advance to obtain an artificial neural network-based prediction model of the city A; A historical climate sample data, determine whether there is a sample filtering rule for the city A, if not, obtain the default historical climate sample data of the city A according to the specific prediction time interval, and use as input data of the artificial neural network; if yes, According to the specific prediction time interval and the sample filtering rule, the filtered historical climate sample data of the city A is obtained and used as the input data of the artificial neural network; then, the historical climate sample data is trained in the artificial neural network for machine learning. The output data of the artificial neural network, the output data is the air quality index learning result; then, comparing the historical air quality index in the input data of the artificial neural network with the error value of the air quality index learning result of the output data, if the error value is greater than a predetermined threshold Continue to perform machine learning; When the error value is less than the predetermined threshold, the machine learning is completed, and the inter-node weight of the city A in the predicted time interval is output and stored, thereby establishing a prediction model of the city A in the predicted time interval. In the forecasting phase, the weather data of city A is first read, and the prediction mode of matching city A is loaded according to the current time. Type, and the weather data of city A as the input data of the artificial neural network, through the artificial neural network for processing, obtain the output data of the artificial neural network, that is, predict the AQI value, and then use the newly acquired AQI to perform the operation cyclically. step. For example, as shown in Table 2, the first round of calculations, according to today's real maximum temperature, minimum temperature, relative humidity, maximum wind speed, maximum wind speed and direction, AQI, and tomorrow, January 15, 2015 The predicted maximum temperature, minimum temperature, relative humidity, maximum wind speed, and maximum wind speed and direction are determined by the artificial neural network-based prediction model to determine the predicted AQI for January 15th, 2015; the second round of calculation process, using 2015 1 The predicted maximum temperature, minimum temperature, relative humidity, maximum wind speed, maximum wind speed and direction, and predicted AQI on January 15, 2015, on January 15 and January 16, 2015, were determined by calculation on January 16, 2015. The predicted AQI; such an operation can be used to determine the AQI prediction results from 15 to 22 days.

图4为本发明中预测空气质量指数的装置一个实施例的结构示意图。Figure 4 is a schematic view showing the structure of an apparatus for predicting an air quality index in the present invention.

所述装置可为于计算机设备，也可为计算机设备中的功能模块；该装置包括模型获取模块410、气象数据获取模块420和指数确定模块430。The device may be a computer device or a functional module in the computer device; the device includes a model acquisition module 410, a weather data acquisition module 420, and an index determination module 430.

首先，模型获取模块410根据待预测地理位置及当前时间，获取相应的预测模型；随后，气象数据获取模块420获取待预测地理位置的气象数据；接着，指数确定模块430根据气象数据，并基于预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。First, the model obtaining module 410 acquires a corresponding prediction model according to the geographic location to be predicted and the current time; subsequently, the meteorological data obtaining module 420 acquires meteorological data of the geographic location to be predicted; then, the index determining module 430 is based on the meteorological data and is based on the prediction The model performs an arithmetic process to determine an air quality index of the to-be-predicted geographic location at the time to be predicted.

本发明的实施例中，可针对每一地理位置在当前时间下均建立相应的预测模型；在执行预测时，首先，获取待预测地理位置的气象数据，该气象数据可从气象部门发布的天气预报等公共信息中获取；随后，根据气象数据，基于相应的预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。本发明相对于现有技术的优势在于：首先，本发明仅需较少的且易于获取的气象信息即可完成未来空气质量指数的预测，由于本发明的方案不需要获取现有技术中大气环流模型所需的复杂大气环流参数，而仅需易于获得的基本气象数据，如历史每日温度、湿度、气压、风力、风向、AQI数值等，即可完成预测，因此任何单位和个人都可以利用气象部门发布的公共气象信息基于本发明所述的技术方案进行AQI预测，同时，基于预测模型，如基于人工神经网络的预测模型的计算量较小，无需高性能硬件设备的高成本投入，降低了预测的实现难度和进入门槛。另外，由于本发明针对每一地理位置在不同的预测时间区间下均建立相应的预测模型，从而可以更好地反映本地区的气候特点，相对于现有技术的大气环流模型有着更好的区域适应性，为确定精确度较高的AQI提供有力保证。In an embodiment of the present invention, a corresponding prediction model may be established for each geographic location at a current time; when performing prediction, first, acquiring meteorological data of a geographic location to be predicted, the weather data may be released from weather by the meteorological department. Obtaining in public information such as forecasting; subsequently, based on the meteorological data, performing arithmetic processing based on the corresponding prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted. The advantages of the present invention over the prior art are: First, the present invention requires only less and easily accessible meteorological information to predict future air quality indices, since the solution of the present invention does not require acquisition of atmospheric circulation in the prior art. The complex atmospheric circulation parameters required by the model, and only basic weather data that are easy to obtain, such as historical daily temperature, humidity, air pressure, wind, wind direction, AQI values, etc., can be used to predict, so that any unit and individual can use The public meteorological information issued by the meteorological department performs AQI prediction based on the technical solution described in the present invention, and at the same time, based on the prediction model, such as the artificial neural network-based prediction model, the calculation amount is small, and high-cost input of high-performance hardware equipment is not required, and the reduction is required. The difficulty of implementation and the threshold for entry. In addition, since the present invention establishes a corresponding prediction model for each geographic location in different prediction time intervals, thereby better reflecting the climate characteristics of the region, there is a better region than the prior art atmospheric circulation model. Adaptability provides a strong guarantee for determining AQI with higher accuracy.

首先，模型获取模块410根据待预测地理位置及当前时间，获取相应的预测模型。First, the model obtaining module 410 acquires a corresponding prediction model according to the geographic location to be predicted and the current time.

当建立基于人工神经网络的预测模型的过程为实时建立时，模型获取模块410 (参照图4)包括预测时间确定单元(图中未示出)、第一指标确定单元(图中未示出)、第一样本获取单元(图中未示出)及第一模型确定单元(图中未示出)；首先，预测时间确定单元确定当前时间所在的预测时间区间；第一指标确定单元确定与待预测地理位置及所述预测时间区间相匹配的气象指标；随后，第一样本获取单元根据待预测地理位置及所述预测时间区间，获取气象指标下的历史气象样本数据；接着，第一模型确定单元根据历史气象样本数据进行机器学习，确定与待预测地理位置及当前时间相对应的预测模型；其中，预测时间确定单元、第一指标确定单元、第一样本获取单元及第一模型确定单元所执行的实时建立预测模型的方式与预先建立预测模型的方式相似，在此参考以下述实施例的具体实施方式。When the process of establishing an artificial neural network based prediction model is established in real time, the model acquisition module 410 (refer to FIG. 4) including a prediction time determining unit (not shown in the drawing), a first index determining unit (not shown in the drawing), a first sample acquiring unit (not shown in the drawing), and a first model determining unit (not shown in the figure); first, the predicted time determining unit determines a predicted time interval in which the current time is located; the first index determining unit determines a weather indicator that matches the predicted geographical position and the predicted time interval; subsequently, the first The sample obtaining unit acquires historical weather sample data under the meteorological indicator according to the geographical position to be predicted and the predicted time interval; then, the first model determining unit performs machine learning according to the historical weather sample data, determines the predicted geographical position and current a time-corresponding prediction model; wherein, the manner in which the prediction time determination unit, the first indicator determination unit, the first sample acquisition unit, and the first model determination unit perform the real-time establishment of the prediction model is similar to the manner in which the prediction model is pre-established, Reference is made herein to the specific embodiments of the following examples.

优选地(参照图4)，该装置还包括第一判断模块(图中未示出)，第一判断模块判断待预测地理位置在预测时间区间内是否存在样本过滤规则；若存在，第一样本获取单元根据待预测地理位置及预测时间区间，基于样本过滤规则，获取气象指标下的历史气象样本数据。Preferably, (refer to FIG. 4), the apparatus further includes a first determining module (not shown), and the first determining module determines whether there is a sample filtering rule in the predicted time interval in the predicted time interval; if present, the first The acquiring unit obtains historical meteorological sample data under the meteorological indicator based on the sample filtering rule according to the geographical position to be predicted and the predicted time interval.

当建立基于人工神经网络的预测模型的过程为预先建立时，模型获取模块410(参照图4)包括预测时间确定单元(图中未示出)和第二模型获取单元(图中未示出)；预测时间确定单元确定当前时间所在的预测时间区间；随后，第二模型获取单元根据待预测地理位置及预测时间区间，在预测模型库中进行匹配查询，获取与待预测地理位置及预测时间区间相对应的预测模型。When the process of establishing the artificial neural network-based prediction model is pre-established, the model acquisition module 410 (refer to FIG. 4) includes a prediction time determination unit (not shown) and a second model acquisition unit (not shown) The prediction time determining unit determines a prediction time interval in which the current time is located; subsequently, the second model obtaining unit performs a matching query in the prediction model library according to the to-be-predicted geographic location and the prediction time interval, and acquires the predicted geographic location and the predicted time interval. Corresponding prediction model.

优选地，预先建立各个地理位置在不同预测时间区间内相应的基于人工神经网络的预测模型，将预测模型存储于预测模型库，并保存地理位置、预测时间区间及预测模型之间的对应关系，如将三者之间的对应关系保存于模型对应列表中，以供查询。Preferably, the artificial neural network-based prediction model of each geographic location in different prediction time intervals is pre-established, the prediction model is stored in the prediction model library, and the correspondence between the geographic location, the prediction time interval, and the prediction model is saved. For example, the correspondence between the three is stored in the model correspondence list for query.

如图5所示，在一优选实施例中，预测空气质量指数的装置还包括第二指标确定模块540、第二样本获取模块550、模型预建模块560和存储模块570；第二指标确定模块540确定与待预测地理位置及预测时间区间相匹配的气象指标；第二样本获取模块550根据待预测地理位置及预测时间区间，获取气象指标下的历史气象样本数据；随后，模型预建模块560根据历史气象样本数据进行机器学习，确定与待预测地理位置及预测时间区间相对应的预测模型；存储模块570将与待预测地理位置及预测时间区间相对应的预测模型保存至预测模型库。As shown in FIG. 5, in a preferred embodiment, the apparatus for predicting an air quality index further includes a second indicator determining module 540, a second sample obtaining module 550, a model pre-built module 560, and a storage module 570; and a second indicator determining module 540: determining a meteorological indicator that matches the predicted geographic location and the predicted time interval; the second sample obtaining module 550 acquires historical meteorological sample data under the meteorological indicator according to the predicted geographic location and the predicted time interval; and then, the model pre-built module 560 The machine learning is performed according to the historical weather sample data, and the prediction model corresponding to the geographical position to be predicted and the predicted time interval is determined; the storage module 570 saves the prediction model corresponding to the geographical position to be predicted and the predicted time interval to the prediction model library.

优选地，该装置还包括更新模块，当预测模型库已存在与待预测地理位置及预测时间区间相对应的预测模型时，更新模块中，用最新的与待预测地理位置及预测时间区间相对应的预测模型替换先前已存在的与待预测地理位置及预测时间区间相对应的预测模型。 Preferably, the apparatus further includes an update module, wherein when the prediction model library already has a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval, the update module corresponds to the latest predicted and predicted time interval and the predicted time interval. The prediction model replaces the previously existing prediction model corresponding to the geographic location to be predicted and the prediction time interval.

第二指标确定模块540确定与待预测地理位置及预测时间区间相匹配的气象指标；The second indicator determining module 540 determines a weather indicator that matches the geographic location to be predicted and the predicted time interval;

例如，待预测地理位置为北京地区，预测时间区间为“1-3月”，在上表1中进行查询匹配，确定与“北京地区”及预测时间区间“1-3月”相匹配的气象指标包括日最高温度、日最低温度、日平均湿度、日主导风向的最大风力以及AQI。For example, the predicted geographic location is Beijing, and the forecast time interval is “1-3 months”. In the above table 1, the query is matched to determine the weather that matches the “Beijing area” and the forecast time interval “1-3 months”. Indicators include daily maximum temperature, daily minimum temperature, daily average humidity, maximum wind force for dominant wind direction, and AQI.

随后，第二样本获取模块550根据待预测地理位置及预测时间区间，获取气象指标下的历史气象样本数据。Subsequently, the second sample obtaining module 550 acquires historical meteorological sample data under the meteorological indicator according to the predicted geographical position and the predicted time interval.

优选地(参照图5)，该装置还包括第二判断模块(图中未示出)；判断模块判断待预测地理位置在预测时间区间内是否存在样本过滤规则；若存在，则第二样本获取模块550根据待预测地理位置及预测时间区间，基于样本过滤规则，获取气象指标下的历史气象样本数据。Preferably (refer to FIG. 5), the apparatus further includes a second determining module (not shown); the determining module determines whether there is a sample filtering rule in the predicted time interval in the predicted time interval; if present, the second sample is obtained The module 550 obtains historical meteorological sample data under the meteorological indicator based on the sample filtering rule according to the geographic location to be predicted and the predicted time interval.

接着，模型预建模块560根据历史气象样本数据进行机器学习，确定与待预测地理位置及预测时间区间相对应的预测模型。Next, the model pre-build module 560 performs machine learning based on the historical weather sample data to determine a prediction model corresponding to the predicted geographic location and the predicted time interval.

具体地，模型预建模块560(参照图5)包括权重确定单元(图中未示出)和模型建立单元(图中未示出)；首先，权重确定单元根据历史气象样本数据，基于人工神经网络进行机器学习，确定与待预测地理位置及预测时间区间相对应的人工神经网络的节点间权重；随后，模型建立单元根据人工神经网络的节点间权重，建立相对应的预测模型。Specifically, the model pre-build module 560 (refer to FIG. 5) includes a weight determination unit (not shown in the figure) and a model establishment unit (not shown); first, the weight determination unit is based on the historical weather sample data, based on the artificial nerve The network performs machine learning to determine the inter-node weights of the artificial neural network corresponding to the predicted geographic location and the predicted time interval. Subsequently, the model establishing unit establishes a corresponding prediction model according to the inter-node weights of the artificial neural network.

具体地，将历史气象样本数据作为人工神经网络的输入数据，基于人工神经网络进行机器学习，确定机器学习结束后人工神经网络的节点间权重；随后，根据人工神经网络的节点间权重，建立相对应的预测模型。Specifically, the historical meteorological sample data is used as the input data of the artificial neural network, and the machine learning is performed based on the artificial neural network to determine the inter-node weight of the artificial neural network after the end of the machine learning; The weights between the nodes of the neural network establish a corresponding prediction model.

其中，权重确定单元根据历史气象样本数据，基于人工神经网络进行机器学习，确定人工神经网络的空气质量指数学习结果；随后，计算空气质量指数学习结果与历史气象样本数据中的空气质量指数的误差值；接着，当误差值小于预定误差阈值时，提取人工神经网络的节点间权重。The weight determining unit determines the air quality index learning result of the artificial neural network based on the historical weather sample data based on the artificial neural network, and then calculates the error of the air quality index in the air quality index learning result and the historical weather sample data. Value; then, when the error value is less than the predetermined error threshold, the inter-node weights of the artificial neural network are extracted.

具体地，将历史气象样本数据作为人工神经网络的输入数据，基于人工神经网络进行机器学习，获取人工神经网络的输出数据，即空气质量指数学习结果；接着，计算输出数据与输入数据中的空气质量指数的误差值，即空气质量指数学习结果与历史气象样本数据中的空气质量指数的误差值；当误差值小于预定误差阈值时，确定机器学习结束，提取人工神经网络的节点间权重。Specifically, the historical meteorological sample data is used as the input data of the artificial neural network, and the machine learning is performed based on the artificial neural network to obtain the output data of the artificial neural network, that is, the air quality index learning result; then, the air in the output data and the input data is calculated. The error value of the quality index, that is, the error value of the air quality index learning result and the air quality index in the historical weather sample data; when the error value is less than the predetermined error threshold, it is determined that the machine learning ends, and the inter-node weight of the artificial neural network is extracted.

存储模块570将与待预测地理位置及预测时间区间相对应的预测模型保存至预测模型库。The storage module 570 saves the prediction model corresponding to the predicted geographic location and the predicted time interval to the prediction model library.

优选地，当预测模型库已存在与待预测地理位置及预测时间区间相对应的预测模型时更新模块用最新的与待预测地理位置及预测时间区间相对应的预测模型替换先前已存在的与待预测地理位置及预测时间区间相对应的预测模型。Preferably, when the prediction model library already has a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval, the update module replaces the previously existing and existing with the latest prediction model corresponding to the predicted geographic location and the predicted time interval. A prediction model that predicts the geographic location and the predicted time interval.

参照图4，气象数据获取模块420获取待预测地理位置的气象数据。Referring to FIG. 4, the weather data acquisition module 420 acquires weather data of a geographic location to be predicted.

具体地，气象数据获取模块420包括第三指标确定单元(图中未示出)和数据确定单元(图中未示出)；第三指标确定单元确定与待预测地理位置及当前时间相匹配的气象指标；数据确定单元根据待预测地理位置，确定气象指标下的气象数据。Specifically, the meteorological data obtaining module 420 includes a third indicator determining unit (not shown in the figure) and a data determining unit (not shown); the third indicator determining unit determines the geographic location to be predicted and the current time. Meteorological indicator; the data determination unit determines the meteorological data under the meteorological indicator according to the geographical position to be predicted.

具体地，首先，根据上表1，确定当前时间对应的预测时间区间，接着，确定与待预测地理位置及预测时间区间相匹配的气象指标；随后，根据待预测地理位置，确定气象指标下的气象数据。 Specifically, first, according to the above Table 1, determining a predicted time interval corresponding to the current time, and then determining a weather indicator that matches the to-be-predicted geographical position and the predicted time interval; and subsequently, determining the meteorological indicator according to the geographical position to be predicted Meteorological data.

例如，待预测时间为明日；则气象数据可包括今日真实的空气质量指数、温度信息、湿度信息、风向信息、风力信息、气压信息、降雨量信息、露点信息等，以及明日预测的温度信息、湿度信息、风向信息、风力信息、气压信息、降雨量信息、露点信息等。For example, the time to be predicted is tomorrow; the weather data may include today's real air quality index, temperature information, humidity information, wind direction information, wind information, air pressure information, rainfall information, dew point information, etc., as well as temperature information predicted tomorrow, Humidity information, wind direction information, wind information, air pressure information, rainfall information, dew point information, etc.

指数确定模块430根据气象数据，基于预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。The index determination module 430 performs an arithmetic process based on the prediction model based on the meteorological data to determine an air quality index of the to-be-predicted geographic location at the time to be predicted.

例如，如上表2所示，2015年1月14日是今日北京地区来自实际观测的真实气象数据，2015年1月15日及以后日期是来自天气预报的预测气象数据。For example, as shown in Table 2 above, January 14, 2015 is the actual meteorological data from actual observations in Beijing today. The date of January 15, 2015 and beyond is the predicted meteorological data from the weather forecast.

本技术领域技术人员可以理解，本发明包括涉及用于执行本申请中所述操作中的一项或多项的设备。这些设备可以为所需的目的而专门设计和制造，或者也可以包括通用计算机中的已知设备。这些设备具有存储在其内的计算机程序，这些计算机程序选择性地激活或重构。这样的计算机程序可以被存储在设备(例如，计算机)可读介质中或者存储在适于存储电子指令并分别耦联到总线的任何类型的介质中，所述计算机可读介质包括但不限于任何类型的盘(包括软盘、硬盘、光盘、CD-ROM、和磁光盘)、ROM(Read-Only Memory，只读存储器)、RAM(Random Access Memory，随即存储器)、EPROM(Erasable Programmable Read-Only Memory，可擦写可编程只读存储器)、EEPROM(Electrically Erasable Programmable Read-Only Memory，电可擦可编程只读存储器)、闪存、磁性卡片或光线卡片。也就是，可读介质包括由设备(例如，计算机)以能够读的形式存储或传输信息的任何介质。Those skilled in the art will appreciate that the present invention includes apparatus related to performing one or more of the operations described herein. These devices may be specially designed and manufactured for the required purposes, or may also include known devices in a general purpose computer. These devices have computer programs stored therein that are selectively activated or reconfigured. Such computer programs may be stored in a device (eg, computer) readable medium or in any type of medium suitable for storing electronic instructions and coupled to a bus, respectively, including but not limited to any Types of disks (including floppy disks, hard disks, optical disks, CD-ROMs, and magneto-optical disks), ROM (Read-Only Memory), RAM (Random Access Memory), EPROM (Erasable Programmable Read-Only Memory) , EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, magnetic card or light card. That is, a readable medium includes any medium that is stored or transmitted by a device (eg, a computer) in a readable form.

本技术领域技术人员可以理解，可以用计算机程序指令来实现这些结构图和/或框图和/或流图中的每个框以及这些结构图和/或框图和/或流图中的框的组合。本技术领域技术人员可以理解，可以将这些计算机程序指令提供给通用计算机、专业计算机或其他可编程数据处理方法的处理器来实现，从而通过计算机或其他可编程数据处理方法的处理器来执行本发明公开的结构图和/或框图和/或流图的框或多个框中指定的方案。Those skilled in the art will appreciate that each block of the block diagrams and/or block diagrams and/or flow diagrams and combinations of blocks in the block diagrams and/or block diagrams and/or flow diagrams can be implemented by computer program instructions. . This technology Those skilled in the art will appreciate that these computer program instructions can be implemented by a general purpose computer, a professional computer, or other programmable data processing method processor to perform the present disclosure by a processor of a computer or other programmable data processing method. The structure diagram and / or block diagram and / or flow diagram of the box or multiple boxes specified in the program.

本发明的实施例还提出了一种程序，包括可读代码，当所述可读代码在计算设备上运行时，导致所述计算设备执行本发明实施例所述的预测空气质量指数的方法。本发明的实施例还提出了一种可读介质，其中存储了上述程序。Embodiments of the present invention also provide a program comprising readable code that, when executed on a computing device, causes the computing device to perform the method of predicting an air quality index as described in embodiments of the present invention. Embodiments of the present invention also propose a readable medium in which the above program is stored.

例如，图6示出了可以实现根据本发明的实现预测空气质量指数的计算设备。该计算设备传统上包括处理器610和以存储器620形式的程序产品或者可读介质。存储器620可以是诸如闪存、EEPROM(电可擦除可编程只读存储器)、EPROM或者ROM之类的电子存储器。存储器620具有用于执行上述方法中的任何方法步骤的程序代码631的存储空间630。例如，用于程序代码的存储空间630可以包括分别用于实现上面的方法中的各种步骤的各个程序代码631。这些程序代码可以从一个或者多个程序产品中读出或者写入到这一个或者多个程序产品中。这些程序产品包括诸如存储卡之类的程序代码载体。这样的程序产品通常为如参考图7所述的便携式或者固定存储单元。该存储单元可以具有与图6的计算设备中的存储器620类似布置的存储段、存储空间等。程序代码可以例如以适当形式进行压缩。通常，存储单元包括可读代码631’，即可以由例如诸如610之类的处理器读取的代码，这些代码当由计算设备运行时，导致该计算设备执行上面所描述的方法中的各个步骤。For example, Figure 6 illustrates a computing device that can implement a predictive air quality index in accordance with the present invention. The computing device conventionally includes a processor 610 and a program product or readable medium in the form of a memory 620. Memory 620 can be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, or ROM. Memory 620 has a memory space 630 for program code 631 for performing any of the method steps described above. For example, storage space 630 for program code may include various program code 631 for implementing various steps in the above methods, respectively. These program codes can be read from or written to one or more program products. These program products include program code carriers such as memory cards. Such a program product is typically a portable or fixed storage unit as described with reference to FIG. The storage unit may have storage segments, storage spaces, and the like that are similarly arranged to memory 620 in the computing device of FIG. The program code can be compressed, for example, in an appropriate form. Typically, the storage unit includes readable code 631', ie, code that can be read by a processor, such as 610, which, when executed by a computing device, causes the computing device to perform various steps in the methods described above. .

本技术领域技术人员可以理解，本发明中已经讨论过的各种操作、方法、流程中的步骤、措施、方案可以被交替、更改、组合或删除。进一步地，具有本发明中已经讨论过的各种操作、方法、流程中的其他步骤、措施、方案也可以被交替、更改、重排、分解、组合或删除。进一步地，现有技术中的具有与本发明中公开的各种操作、方法、流程中的步骤、措施、方案也可以被交替、更改、重排、分解、组合或删除。Those skilled in the art can understand that the steps, measures, and solutions in the various operations, methods, and processes that have been discussed in the present invention may be alternated, changed, combined, or deleted. Further, other steps, measures, and schemes of the various operations, methods, and processes that have been discussed in the present invention may be alternated, modified, rearranged, decomposed, combined, or deleted. Further, the steps, measures, and solutions in the prior art having various operations, methods, and processes disclosed in the present invention may also be alternated, changed, rearranged, decomposed, combined, or deleted.

以上所述仅是本发明的部分实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。 The above is only a part of the embodiments of the present invention, and it should be noted that those skilled in the art can also make several improvements and retouchings without departing from the principles of the present invention. It should be considered as the scope of protection of the present invention.

Claims

一种预测空气质量指数的方法，包括：A method of predicting an air quality index, including:

根据待预测地理位置及当前时间，获取相应的预测模型；Obtain a corresponding prediction model according to the geographical location to be predicted and the current time;

获取所述待预测地理位置的气象数据；以及Obtaining meteorological data of the geographic location to be predicted;

根据所述气象数据，并基于所述预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。Based on the meteorological data, and performing an arithmetic process based on the prediction model, determining an air quality index of the to-be-predicted geographic location at the time to be predicted.
根据权利要求1所述的预测空气质量指数的方法，其特征在于，所述根据待预测地理位置及当前时间，获取相应的预测模型的步骤进一步包括：The method for predicting an air quality index according to claim 1, wherein the step of obtaining a corresponding prediction model according to the geographic location to be predicted and the current time further comprises:

确定当前时间所在的预测时间区间；Determining the forecast time interval in which the current time is located;

确定与待预测地理位置及预测时间区间相匹配的气象指标；Determining meteorological indicators that match the geographic location to be predicted and the predicted time interval;

根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据；以及Obtaining historical meteorological sample data under the meteorological indicator according to the predicted geographic location and the predicted time interval;

根据所述历史气象样本数据进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的预测模型。Performing machine learning based on the historical weather sample data, and determining a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval.
根据权利要求2所述的预测空气质量指数的方法，该方法还包括：The method of predicting an air quality index according to claim 2, further comprising:

判断所述待预测地理位置在所述预测时间区间内是否存在样本过滤规则；Determining whether there is a sample filtering rule in the predicted time interval of the to-be-predicted geographic location;

若存在，则所述根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据的步骤进一步包括：If yes, the step of acquiring historical meteorological sample data under the meteorological indicator according to the to-be-predicted geographic location and the predicted time interval further includes:

根据所述待预测地理位置及所述预测时间区间，基于所述样本过滤规则，获取所述气象指标下的历史气象样本数据。Obtaining historical weather sample data under the meteorological indicator based on the sample filtering rule according to the to-be-predicted geographic location and the predicted time interval.
根据权利要求1所述的预测空气质量指数的方法，其中，所述根据待预测地理位置及当前时间，获取相应的预测模型的步骤进一步包括：The method for predicting an air quality index according to claim 1, wherein the step of acquiring a corresponding prediction model according to the geographic location to be predicted and the current time further comprises:

确定当前时间所在的预测时间区间；以及Determining the forecast time interval in which the current time is located;

根据所述待预测地理位置及所述预测时间区间，在预测模型库中进行匹配查询，获取与所述待预测地理位置及所述当前时间相对应的预测模型。And performing a matching query in the prediction model library according to the to-be-predicted geographic location and the prediction time interval, and acquiring a prediction model corresponding to the to-be-predicted geographic location and the current time.
根据权利要求4所述的预测空气质量指数的方法，其中，所述根据待预测地理位置及预测时间区间，在预测模型库预测模型库中进行匹配查询，获取与待预测地理位置及当前时间相对应的预测模型的步骤之前，该方法还包括：The method for predicting an air quality index according to claim 4, wherein the matching query is performed in a predictive model library predictive model library according to a geographic location to be predicted and a predicted time interval, and the obtained geographic location and current time are obtained. Before the steps of the corresponding prediction model, the method further includes:

确定与待预测地理位置及预测时间区间相匹配的气象指标；Determining meteorological indicators that match the geographic location to be predicted and the predicted time interval;

根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据；Obtaining historical meteorological sample data under the meteorological indicator according to the predicted geographic location and the predicted time interval;

根据所述历史气象样本数据进行机器学习，以确定与所述待预测地理位置及所述预测时间区间相对应的预测模型；以及Performing machine learning based on the historical weather sample data to determine a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval;

将与所述待预测地理位置及所述预测时间区间相对应的预测模型保存至所述预测模型库。 Predicting a model corresponding to the to-be-predicted geographic location and the predicted time interval is saved to the predictive model library.
根据权利要求5所述的预测空气质量指数的方法，其中，该方法还包括：The method of predicting an air quality index according to claim 5, wherein the method further comprises:

当所述预测模型库已存在与所述待预测地理位置及所述预测时间区间相对应的预测模型时，用最新的与所述待预测地理位置及所述预测时间区间相对应的预测模型替换先前已存在的与所述待预测地理位置及所述预测时间区间相对应的预测模型。When the prediction model library already has a prediction model corresponding to the to-be-predicted geographic location and the prediction time interval, replacing with the latest prediction model corresponding to the to-be-predicted geographic location and the predicted time interval A prediction model that previously existed corresponding to the to-be-predicted geographic location and the predicted time interval.
根据权利要求5或6所述的预测空气质量指数的方法，该方法还包括：The method of predicting an air quality index according to claim 5 or 6, further comprising:

判断所述待预测地理位置在所述预测时间区间内是否存在样本过滤规则；Determining whether there is a sample filtering rule in the predicted time interval of the to-be-predicted geographic location;

若存在，则所述根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据的步骤进一步包括：If yes, the step of acquiring historical meteorological sample data under the meteorological indicator according to the to-be-predicted geographic location and the predicted time interval further includes:

根据所述待预测地理位置及所述预测时间区间，基于所述样本过滤规则，获取所述气象指标下的历史气象样本数据。Obtaining historical weather sample data under the meteorological indicator based on the sample filtering rule according to the to-be-predicted geographic location and the predicted time interval.
根据权利要求1-7中任一项所述的预测空气质量指数的方法，其特征在于，所述预测模型为“基于人工神经网络”的预测模型。The method of predicting an air quality index according to any one of claims 1 to 7, wherein the prediction model is a prediction model based on an artificial neural network.
根据权利要求2-8中任一项所述的预测空气质量指数的方法，其中，所述根据所述历史气象样本数据进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的预测模型的步骤进一步包括：The method of predicting an air quality index according to any one of claims 2-8, wherein said performing machine learning based on said historical weather sample data determines that said predicted geographic location and said predicted time interval are The steps of the corresponding prediction model further include:

根据所述历史气象样本数据，基于人工神经网络进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的人工神经网络的节点间权重；Performing machine learning based on the artificial neural network according to the historical weather sample data, and determining an inter-node weight of the artificial neural network corresponding to the to-be-predicted geographic location and the predicted time interval;

根据所述人工神经网络的节点间权重，建立相对应的预测模型。Corresponding prediction models are established according to the weights between nodes of the artificial neural network.
根据权利要求9所述的预测空气质量指数的方法，其中，所述根据所述历史气象样本数据，基于人工神经网络进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的人工神经网络的节点间权重的步骤进一步包括：The method of predicting an air quality index according to claim 9, wherein said determining, based on said historical weather sample data, based on an artificial neural network, machine learning, corresponding to said predicted geographical position and said predicted time interval The steps of weighting between nodes of the artificial neural network further include:

根据所述历史气象样本数据，基于人工神经网络进行机器学习，确定所述人工神经网络的空气质量指数学习结果；Determining an air quality index learning result of the artificial neural network according to the historical weather sample data, performing machine learning based on an artificial neural network;

计算所述空气质量指数学习结果与所述历史气象样本数据中的历史空气质量指数的误差值；Calculating an error value of the air quality index learning result and a historical air quality index in the historical weather sample data;

当所述误差值小于预定误差阈值时，提取所述人工神经网络的节点间权重。When the error value is less than the predetermined error threshold, the inter-node weight of the artificial neural network is extracted.
根据权利要求1-10中任一项所述的预测空气质量指数的方法，其中，所述获取待预测地理位置的气象数据的步骤进一步包括：The method of predicting an air quality index according to any one of claims 1 to 10, wherein the step of acquiring weather data of a geographical position to be predicted further comprises:

确定与待预测地理位置及当前时间相匹配的气象指标；Determining meteorological indicators that match the geographic location to be predicted and the current time;

根据所述待预测地理位置，确定所述气象指标下的气象数据。Meteorological data under the meteorological indicator is determined according to the predicted geographic location.
根据权利要求1-11中任一项所述的预测空气质量指数的方法，其中，所述气象指标包括以下至少任一项：The method of predicting an air quality index according to any one of claims 1 to 11, wherein the meteorological indicator comprises at least one of the following:

温度指标；湿度指标；风向指标；风力指标；气压指标；降雨量指标；露点指标；空气质量指数指标。Temperature index; humidity index; wind direction index; wind index; air pressure index; rainfall index; dew point index; air quality index.
根据权利要求1-12中任一项所述的预测空气质量指数的方法，其中，所述待预测时间为明日，所述气象数据包括以下至少任一项：A method of predicting an air quality index according to any one of claims 1 to 12, wherein said The forecast time is tomorrow, and the meteorological data includes at least one of the following:

今日的气象数据及明日的预测气象数据；Today's weather data and forecast weather data for tomorrow;

前预定天数的真实气象数据，今日的气象数据及明日的预测气象数据；Real weather data for the previous scheduled days, today's weather data and forecast weather data for tomorrow;

其中，今日的气象数据包括以下至少任一项：Among them, today's weather data includes at least one of the following:

今日的真实气象数据；今日的预测气象数据；Real weather data today; today's forecast meteorological data;

其中，预测气象数据包括以下至少任一项：Among them, the predicted meteorological data includes at least one of the following:

温度信息；湿度信息；风向信息；风力信息；气压信息；降雨量信息；露点信息。Temperature information; humidity information; wind direction information; wind information; air pressure information; rainfall information; dew point information.
一种预测空气质量指数的装置，包括：A device for predicting an air quality index, comprising:

模型获取模块，用于根据待预测地理位置及当前时间，获取相应的预测模型；a model obtaining module, configured to acquire a corresponding prediction model according to the geographic location to be predicted and the current time;

气象数据获取模块，用于获取所述待预测地理位置的气象数据；以及a meteorological data acquisition module, configured to acquire weather data of the geographic location to be predicted;

指数确定模块，用于根据所述气象数据，并基于所述预测模型进行运算处理，确定待预测地理位置在待预测时间的空气质量指数。The index determining module is configured to perform, according to the meteorological data, an operation process based on the prediction model to determine an air quality index of the to-be-predicted geographic location at the time to be predicted.
根据权利要求14所述的预测空气质量指数的装置，其中，所述模型获取模块进一步包括：The apparatus for predicting an air quality index according to claim 14, wherein the model acquisition module further comprises:

预测时间确定单元，用于确定当前时间所在的预测时间区间；a prediction time determining unit, configured to determine a predicted time interval in which the current time is located;

第一指标确定单元，用于确定与待预测地理位置及预测时间区间相匹配的气象指标；a first indicator determining unit, configured to determine a weather indicator that matches a geographic location to be predicted and a predicted time interval;

第一样本获取单元，用于根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据；以及a first sample obtaining unit, configured to acquire historical meteorological sample data under the meteorological indicator according to the to-be-predicted geographic location and the predicted time interval;

第一模型确定单元，用于根据所述历史气象样本数据进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的预测模型。And a first model determining unit, configured to perform machine learning according to the historical weather sample data, and determine a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval.
根据权利要求15所述的预测空气质量指数的装置，该装置还包括：The apparatus for predicting an air quality index according to claim 15, further comprising:

第一判断模块，用于判断所述待预测地理位置在所述预测时间区间内是否存在样本过滤规则；a first determining module, configured to determine whether a sample filtering rule exists in the predicted time interval in the predicted time interval;

若存在，所述第一样本获取模块用于根据所述待预测地理位置及所述预测时间区间，基于所述样本过滤规则，获取所述气象指标下的历史气象样本数据。If yes, the first sample obtaining module is configured to acquire historical meteorological sample data under the meteorological indicator based on the sample filtering rule according to the to-be-predicted geographic location and the predicted time interval.
根据权利要求14所述的预测空气质量指数的装置，其中，所述模型获取模块进一步包括：The apparatus for predicting an air quality index according to claim 14, wherein the model acquisition module further comprises:

预测时间确定单元，用于确定当前时间所在的预测时间区间；以及a prediction time determining unit, configured to determine a predicted time interval in which the current time is located;

第二模型获取单元，用于根据所述待预测地理位置及所述预测时间区间，在预测模型库中进行匹配查询，获取与所述待预测地理位置及所述当前时间相对应的预测模型。And a second model acquiring unit, configured to perform a matching query in the prediction model library according to the to-be-predicted geographic location and the prediction time interval, and acquire a prediction model corresponding to the to-be-predicted geographic location and the current time.
根据权利要求17所述的预测空气质量指数的装置，其中，该装置还包括：The apparatus for predicting an air quality index according to claim 17, wherein the apparatus further comprises:

第二指标确定模块，用于确定与所述待预测地理位置及所述预测时间区间相匹配的气象指标；a second indicator determining module, configured to determine that the predicted geographic location and the predicted time interval are matched Meteorological indicators

第二样本获取模块，用于根据所述待预测地理位置及所述预测时间区间，获取所述气象指标下的历史气象样本数据；a second sample obtaining module, configured to acquire historical meteorological sample data under the meteorological indicator according to the geographic location to be predicted and the predicted time interval;

模型预建模块，用于根据所述历史气象样本数据进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的预测模型；以及a model pre-built module, configured to perform machine learning according to the historical weather sample data, and determine a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval;

存储模块，用于将与所述待预测地理位置及所述预测时间区间相对应的预测模型保存至所述预测模型库。And a storage module, configured to save the prediction model corresponding to the to-be-predicted geographic location and the predicted time interval to the prediction model library.
根据权利要求18所述的预测空气质量指数的装置，其中，该装置还包括：The apparatus for predicting an air quality index according to claim 18, wherein the apparatus further comprises:

更新模块，用于当所述预测模型库已存在与所述待预测地理位置及所述预测时间区间相对应的预测模型时，用最新的与所述待预测地理位置及所述预测时间区间相对应的预测模型替换先前已存在的与所述待预测地理位置及所述预测时间区间相对应的预测模型。And an update module, configured to: when the prediction model library has a prediction model corresponding to the to-be-predicted geographic location and the predicted time interval, use the latest predicted geographic location and the predicted time interval The corresponding prediction model replaces the previously existing prediction model corresponding to the to-be-predicted geographic location and the predicted time interval.
根据权利要求18或19所述的预测空气质量指数的装置，该装置还包括：The apparatus for predicting an air quality index according to claim 18 or 19, further comprising:

第二判断模块，用于判断所述待预测地理位置在所述预测时间区间内是否存在样本过滤规则；a second determining module, configured to determine whether a sample filtering rule exists in the predicted time interval in the predicted time interval;

若存在，所述第二样本获取模块用于根据所述待预测地理位置及所述预测时间区间，基于所述样本过滤规则，获取所述气象指标下的历史气象样本数据。If yes, the second sample obtaining module is configured to acquire historical meteorological sample data under the meteorological indicator based on the sample filtering rule according to the to-be-predicted geographic location and the predicted time interval.
根据权利要求14-20中任一项所述的预测空气质量指数的装置，其特征在于，所述预测模型为“基于人工神经网络”的预测模型。The apparatus for predicting an air quality index according to any one of claims 14 to 20, wherein the prediction model is a prediction model based on an artificial neural network.
根据权利要求15-20中任一项所述的预测空气质量指数的装置，其中，所述模型预建模块包括：The apparatus for predicting an air quality index according to any one of claims 15 to 20, wherein the model pre-built module comprises:

权重确定单元，用于根据所述历史气象样本数据，基于人工神经网络进行机器学习，确定与所述待预测地理位置及所述预测时间区间相对应的人工神经网络的节点间权重；以及a weight determining unit, configured to perform machine learning based on the artificial neural network according to the historical weather sample data, and determine an inter-node weight of the artificial neural network corresponding to the to-be-predicted geographic location and the predicted time interval;

模型建立单元，用于根据所述人工神经网络的节点间权重，建立相对应的预测模型。And a model establishing unit, configured to establish a corresponding prediction model according to the weight between nodes of the artificial neural network.
根据权利要求22所述的预测空气质量指数的装置，其中，所述权重确定单元用于根据所述历史气象样本数据，基于人工神经网络进行机器学习，确定所述人工神经网络的学习结果数据；计算所述空气质量指数学习结果与所述历史气象样本数据中的历史空气质量指数的误差值；以及当所述误差值小于预定误差阈值时，提取所述人工神经网络的节点间权重。The apparatus for predicting an air quality index according to claim 22, wherein the weight determining unit is configured to determine, according to the historical weather sample data, machine learning based on an artificial neural network, learning result data of the artificial neural network; Calculating an error value of the air quality index learning result and a historical air quality index in the historical weather sample data; and extracting an inter-node weight of the artificial neural network when the error value is less than a predetermined error threshold.
根据权利要求14-23中任一项所述的预测空气质量指数的装置，其中，所述气象数据获取模块包括：The apparatus for predicting an air quality index according to any one of claims 14 to 23, wherein the weather data acquisition module comprises:

第三指标确定单元，用于确定与待预测地理位置及当前时间相匹配的气象指标；a third indicator determining unit, configured to determine a weather indicator that matches the geographic location to be predicted and the current time;

数据确定单元，用于根据所述待预测地理位置，确定所述气象指标下的气象数据。a data determining unit, configured to determine a meteorological quantity under the meteorological indicator according to the geographic location to be predicted according to.
根据权利要求14-24中任一项所述的预测空气质量指数的方法，其中，所述气象指标包括以下至少任一项：A method of predicting an air quality index according to any one of claims 14 to 24, wherein the meteorological indicator comprises at least one of the following:

温度指标；湿度指标；风向指标；风力指标；气压指标；降雨量指标；露点指标；空气质量指数指标。Temperature index; humidity index; wind direction index; wind index; air pressure index; rainfall index; dew point index; air quality index.
根据权利要求14-25中任一项所述的预测空气质量指数的方法，其中，所述待预测时间为明日，所述气象数据包括以下至少任一项：The method of predicting an air quality index according to any one of claims 14 to 25, wherein the to-be-predicted time is tomorrow, and the meteorological data includes at least one of the following:

今日的气象数据及明日的预测气象数据；Today's weather data and forecast weather data for tomorrow;

前预定天数的真实气象数据，今日的气象数据及明日的预测气象数据；Real weather data for the previous scheduled days, today's weather data and forecast weather data for tomorrow;

其中，今日的气象数据包括以下至少任一项：Among them, today's weather data includes at least one of the following:

今日的真实气象数据；今日的预测气象数据；Real weather data today; today's forecast meteorological data;

其中，预测气象数据包括以下至少任一项：Among them, the predicted meteorological data includes at least one of the following:

温度信息；湿度信息；风向信息；风力信息；气压信息；降雨量信息；露点信息。Temperature information; humidity information; wind direction information; wind information; air pressure information; rainfall information; dew point information.
一种程序，包括可读代码，当所述可读代码在计算设备上运行时，导致所述计算设备执行根据权利要求1-13中的任一个所述的预测空气质量指数的方法。A program comprising readable code that, when executed on a computing device, causes the computing device to perform the method of predicting an air quality index according to any of claims 1-13.
一种可读介质，其中存储了如权利要求27所述的程序。 A readable medium storing the program of claim 27.