WO2019205804A1

WO2019205804A1 - Web page pre-downloading method and device, storage medium and electronic device

Info

Publication number: WO2019205804A1
Application number: PCT/CN2019/076495
Authority: WO
Inventors: 曹刚
Original assignee: 中兴通讯股份有限公司
Priority date: 2018-04-23
Filing date: 2019-02-28
Publication date: 2019-10-31
Also published as: CN110392085A

Abstract

Provided by the present application are a web page pre-downloading method and device, a storage medium and an electronic device. The method comprises: when a terminal is currently connected to a network by means of WiFi and/or the current time reaches a web page pre-download time point, the terminal determining web pages which need to be pre-downloaded according to the content attributes of the web pages and downloading the determined web pages. By employing the present solution, it is possible to determine which web pages need to be pre-downloaded on the basis of the content attributes of web pages when there is WiFi and/or a web page pre-download time point has been reached.

Description

网页预下载方法及装置、存储介质和电子装置Webpage pre-downloading method and device, storage medium and electronic device

相关申请的交叉引用Cross-reference to related applications

本申请基于申请号为201810365665.0、申请日为2018年04月23日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。The present application is filed on the basis of the Chinese Patent Application Serial No. 20, 181, 036, 566, filed on Apr. 23, 2011, the entire disclosure of which is hereby incorporated by reference.

技术领域Technical field

本申请涉及通信领域但不限于通信领域，尤其涉及一种网页预下载方法及装置、存储介质和电子装置。The present application relates to the field of communications, but is not limited to the field of communications, and in particular, to a webpage pre-download method and apparatus, a storage medium, and an electronic device.

背景技术Background technique

随着无线通讯技术的迅猛发展和普及，现在越来越多的用户习惯在移动终端或固定终端上通过浏览器或专用手机、电脑应用(Application，简称为APP)浏览各种新闻资讯。With the rapid development and popularization of wireless communication technologies, more and more users are now accustomed to browsing various news information on mobile terminals or fixed terminals through browsers or dedicated mobile phones and computer applications (Application, referred to as APP).

但是，在浏览新闻资讯的过程中，往往会存在未处于无线保真(Wireless Fidelity，简称为WIFI)覆盖环境导致需要使用移动流量来浏览新闻资讯，或者即使用户愿意使用移动流量来浏览新闻资讯，但是因数据网络信号较弱导致网页打开缓慢的情况，这会极大地影响用户的使用体验。However, in the process of browsing news information, there is often a wireless Fidelity (WIFI) coverage environment that requires the use of mobile traffic to browse news information, or even if users are willing to use mobile traffic to browse news information. However, due to the weak data network signal, the webpage is slowly opened, which greatly affects the user experience.

发明内容Summary of the invention

本申请实施例提供了一种网页预下载方法及装置、存储介质和电子装置。The embodiment of the present application provides a webpage pre-downloading method and device, a storage medium, and an electronic device.

本申请实施例提供了一种网页预下载方法，包括：在终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，所述终端根据网页的内容属性确定需要预下载的网页；所述终端下载所述确定的网页。The embodiment of the present application provides a webpage pre-downloading method, including: when the terminal currently connects to the webpage pre-downloading time point through the wireless fidelity WIFI connection network and/or the current time, the terminal determines the requirement according to the content attribute of the webpage. a pre-downloaded webpage; the terminal downloading the determined webpage.

本申请实施例还提供了一种网页预下载装置，应用于终端，所述装置包括：网页预载决策模块，配置为在所述终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，根据网页的内容属性确定需要预下载的网页；网页下载模块，配置为下载所述确定的网页。The embodiment of the present application further provides a webpage pre-downloading device, which is applied to a terminal, and the device includes: a webpage preloading decision module, configured to reach a webpage by using a wireless fidelity WIFI connection network and/or a current time at the terminal. In the case of pre-downloading the time point, the webpage that needs to be pre-downloaded is determined according to the content attribute of the webpage; the webpage downloading module is configured to download the determined webpage.

本申请实施例还提供了一种存储介质，所述存储介质中存储有计算机程序，其中，所述计算机程序被设置为运行时执行以上所述的方法。The embodiment of the present application further provides a storage medium in which a computer program is stored, wherein the computer program is set to execute the method described above at runtime.

本申请实施例还提供了一种电子装置，包括存储器和处理器，所述存储器中存储有计算机程序，所述处理器被设置为运行所述计算机程序以执行以上所述的方法。Embodiments of the present application also provide an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the method described above.

通过本申请，在终端当前通过WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，所述终端根据网页的内容属性确定需要预下载的网页，并下载所述确定的网页。该方案能够在有WIFI和/或到达网页预下载时间点的情况下，基于网页的内容属性确定哪些网页要进行预下载。由于进行预下载的网页是根据网页的内容属性来确定的，可以实现更加有针对性的网页预下载，在避免下载全部网页导致内存耗尽的情况下，保证用户在后续需要浏览网页时的浏览体验，可以解决相关技术中用户浏览网页需使用移动流量或打开缓慢的问题，提高了用户的体验。In the case that the terminal currently reaches the webpage pre-downloading time point through the WIFI connection network and/or the current time, the terminal determines the webpage that needs to be pre-downloaded according to the content attribute of the webpage, and downloads the determined webpage. The solution can determine which pages are to be pre-downloaded based on the content attributes of the webpage when there is a WIFI and/or a webpage pre-download time point. Since the pre-downloaded webpage is determined according to the content attribute of the webpage, a more targeted pre-downloading of the webpage can be realized, and in the case of avoiding the exhaustion of the memory by downloading all the webpages, the user is required to browse in the subsequent browsing of the webpage. The experience can solve the problem that the user needs to use mobile traffic or open slowly in browsing the webpage in related technologies, thereby improving the user experience.

附图说明DRAWINGS

此处所说明的附图用来提供对本申请的进一步理解，构成本申请的一部分，本申请的示意性实施例及其说明用于解释本申请，并不构成对本申请的不当限定。在附图中：The drawings described herein are intended to provide a further understanding of the present application, and are intended to be a part of this application. In the drawing:

图1是本申请实施例的一种网页预下载方法的移动终端的硬件结构框图；1 is a block diagram showing the hardware structure of a mobile terminal of a webpage pre-download method according to an embodiment of the present application;

图2是根据本申请实施例的网页预下载方法的流程图；2 is a flowchart of a webpage pre-download method according to an embodiment of the present application;

图3是根据本申请实施例的网页预下载装置的结构框图；3 is a structural block diagram of a webpage pre-downloading apparatus according to an embodiment of the present application;

图4是根据本申请实施例的网页预下载装置的第一优选结构框图；4 is a first preferred structural block diagram of a webpage pre-downloading apparatus according to an embodiment of the present application;

图5是根据本申请实施例的网页预下载装置的第二优选结构框图；FIG. 5 is a second preferred structural block diagram of a webpage pre-downloading apparatus according to an embodiment of the present application; FIG.

图6是根据本申请实施例的网页预下载装置的第三优选结构框图；FIG. 6 is a third preferred structural block diagram of a webpage pre-downloading apparatus according to an embodiment of the present application; FIG.

图7是根据本申请实施例的用于实现基于机器学习的网页预加载方法的***的结构框图；7 is a structural block diagram of a system for implementing a machine learning based webpage preloading method according to an embodiment of the present application;

图8是根据本申请实施例的输入向量的构造方法的流程图；FIG. 8 is a flowchart of a method of constructing an input vector according to an embodiment of the present application; FIG.

图9是根据本申请实施例的网页预加载决策子模块74的处理流程图。FIG. 9 is a process flow diagram of a webpage preloading decision sub-module 74 in accordance with an embodiment of the present application.

具体实施方式detailed description

在用户浏览新闻资讯的过程中，往往会存在未处于WIFI等网络覆盖环境导致需要使用移动流量来浏览新闻资讯，或者即使用户愿意使用移动流量来浏览新闻资讯，但是因数据网络信号较弱导致网页打开缓慢的情况。考虑到以上情况，终端用户可能需要手机在家里或办公室有WIFI等网络的时候，或者在自己浏览新闻资讯之前，能自动在后台把自己经常浏览的网站上喜欢的网页提前加载下来，这样到了非WIFI环境下(比如上下班的地铁、公交或班车上)能够无需使用终端SIM卡流量而很快地浏览到这些网页。In the process of browsing news information, users often have to use mobile traffic to browse news information without being in a network coverage environment such as WIFI, or even if users are willing to use mobile traffic to browse news information, but because the data network signal is weak, the webpage is weak. Turn on slowly. In view of the above situation, the end user may need to have a mobile phone at home or in the office with a network such as WIFI, or automatically browse the favorite pages on the website that he frequently browses in the background before browsing the news. In a WIFI environment (such as on a subway, bus or shuttle bus), you can quickly browse to these pages without using the terminal SIM card traffic.

为了满足上述的用户需求，某些新闻客户端APP提供了一种定时后台加载网页的功能，会根据用户的设置定时在一个时间段内提前将本新闻客户端当前主页上所有的网页均保存在终端磁盘上。后续用户在浏览这些网页时可以不需要联网(WIFI或移动流量)，而是直接从终端磁盘上保存的网页数据加载到APP界面上。但是，该方案存在以下问题：In order to meet the above user requirements, some news client APPs provide a function of periodically loading webpages in a timed manner, and all webpages on the current homepage of the news client are saved in advance in a time period according to the setting time of the user. On the terminal disk. Subsequent users can browse the web pages without the need for networking (WIFI or mobile traffic), but load the web page data saved directly from the terminal disk to the APP interface. However, the program has the following problems:

(1)因为该方案要批量后台下载当前新闻客户端主页上所有的网页，这对终端***资源消耗很大(如中央处理器(Central Processing Unit，简称为CPU)和存储空间)，而很多网页并不是用户平常经常喜爱浏览的网页(即缺乏个性化操作)，从而造成***资源浪费。(1) Because the program wants to download all the web pages on the current news client homepage in batches, this is very expensive for the terminal system resources (such as Central Processing Unit (CPU) and storage space), and many web pages. It is not a web page that users often like to browse frequently (that is, lack of personalized operation), resulting in wasted system resources.

(2)因为是通过用户设置的时间定时操作，时间提前或延后可能分别造成最新网页没有及时更新下载或来不及下载的情况。(2) Because it is a time-timed operation set by the user, the advance or delay of the time may cause the latest webpage not to be updated or downloaded in time or downloaded.

(3)因为提前下载的网页都保存在本地磁盘上，加载起来也需要一定时间，同时一些有绝对路径的链接会找不到子资源从而影响网页的显示。(3) Because the web pages downloaded in advance are saved on the local disk, it takes a certain time to load, and some links with absolute paths will not find the sub-resources and affect the display of the web pages.

本申请实施例提供了一种网页预下载的方案。该方案在终端有WIFI网络和/或已到达网页预下载时间点的情况下，终端可以根据网页的内容属性确定需要预下载的网页。例如，这些网页可以是用户平常喜爱浏览的网页，终端可以将需要预下载的网页下载下来。The embodiment of the present application provides a scheme for pre-downloading a webpage. In the case that the terminal has a WIFI network and/or a pre-downloading time point of the webpage has arrived, the terminal may determine a webpage that needs to be pre-downloaded according to the content attribute of the webpage. For example, the webpages may be webpages that users usually like to browse, and the terminal may download webpages that need to be pre-downloaded.

该方案还可以在后台进行预加载(例如，可以包括对网页数据进行解码和解析、布局、GPU绘制等渲染工作，以及将要显示的内容准备好放到缓存中等等为能够正常显示网页所做的准备工作)并保存到内存中，这样，当用户在无WIFI或网络信号较弱等情况下，可以直接从内核中将网页内容切换到前台显示。The solution can also be preloaded in the background (for example, it can include rendering and parsing of webpage data, layout, GPU rendering, etc., and preparing the content to be displayed in the cache, etc. for the normal display of the webpage. Prepare the work) and save it to the memory, so that when the user has no WIFI or the network signal is weak, the content of the webpage can be directly switched from the kernel to the foreground display.

该方案中网页预下载时间点可以是基于用户的平均上网时段确定的，更加符合用户的浏览习惯，能够尽量避免最新网页没有及时更新下载或来不及下载的情况。The pre-downloading time point of the webpage in the solution may be determined based on the average online time of the user, and is more in line with the browsing habit of the user, and can avoid the situation that the latest webpage is not updated or downloaded in time.

根据网页的内容属性确定需要预下载的网页的过程所基于的模型可以是基于用户行为自动调节和学习的，从而能够做到更加精准的网页预加载选择。The model based on determining the content of the webpage that needs to be pre-downloaded according to the content attribute of the webpage may be automatically adjusted and learned based on the user behavior, so that more precise webpage preloading selection can be achieved.

下文中将参考附图并结合实施例来详细说明本申请。需要说明的是，在不冲突的情况下，本申请中的实施例及实施例中的特征可以相互组合。The present application will be described in detail below with reference to the drawings in conjunction with the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict.

需要说明的是，本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。It should be noted that the terms "first", "second" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or order.

本申请实施例一所提供的方法实施例可以在移动终端、计算机终端或者类似的运算装置中执行。以运行在移动终端上为例，图1是本申请实施例的一种网页预下载方法的移动终端的硬件结构框图。如图1所示，移动终端10可以包括一个或多个(图1中仅示出一个)处理器102(处理器102可以包括但不限于微处理器MCU或可编程逻辑器件FPGA等的处理装置)和用于存储数据的存储器104，可选地，上述移动终端还可以包括用于通信功能的传输设备106以及输入输出设备108。本领域普通技术人员可以理解，图1所示的结构仅为示意，其并不对上述移动终端的结构造成限定。例如，移动终端10还可包括比图1中所示更多或者更少的组件，或者具有与图1所示不同的配置。The method embodiment provided in Embodiment 1 of the present application can be executed in a mobile terminal, a computer terminal or the like. Taking a mobile terminal as an example, FIG. 1 is a hardware structural block diagram of a mobile terminal of a webpage pre-downloading method according to an embodiment of the present application. As shown in FIG. 1, mobile terminal 10 may include one or more (only one shown in FIG. 1) processor 102 (processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA. And a memory 104 for storing data, optionally, the above mobile terminal may further include a transmission device 106 for communication functions and an input and output device 108. It will be understood by those skilled in the art that the structure shown in FIG. 1 is merely illustrative, and does not limit the structure of the above mobile terminal. For example, the mobile terminal 10 may also include more or fewer components than those shown in FIG. 1, or have a different configuration than that shown in FIG.

存储器104可用于存储计算机程序，例如，应用软件的软件程序以及模块，如本申请实施例中的网页预下载方法对应的计算机程序，处理器102通过运行存储在存储器104内的计算机程序，从而执行各种功能应用以及数据处理，即实现上述的方法。存储器104可包括高速随机存储器，还可包括非易失性存储器，如一个或者多个磁性存储装置、闪存、或者其他非易失性固态存储器。在一些实例中，存储器104可进一步包括相对于处理器102远程设置的存储器，这些远程存储器可以通过网络连接至移动终端10。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。The memory 104 can be used to store a computer program, for example, a software program and a module of the application software, such as a computer program corresponding to the webpage pre-download method in the embodiment of the present application, and the processor 102 executes by executing a computer program stored in the memory 104. Various functional applications and data processing, that is, the above methods are implemented. Memory 104 may include high speed random access memory, and may also include non-volatile memory such as one or more magnetic storage devices, flash memory, or other non-volatile solid state memory. In some examples, memory 104 may further include memory remotely located relative to processor 102, which may be connected to mobile terminal 10 over a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

传输设备106用于经由一个网络接收或者发送数据。上述的网络具体实例可包括移动终端10的通信供应商提供的无线网络。在一个实例中，传输设备106包括一个网络适配器(Network Interface Controller，简称为NIC)，其可通过基站与其他网络设备相连从而可与互联网进行通讯。在一个实例中，传输设备106可以为射频(Radio Frequency，简称为RF)模块，RF模块配置为通过无线方式与互联网进行通讯。 Transmission device 106 is for receiving or transmitting data via a network. The above-described network specific example may include a wireless network provided by a communication provider of the mobile terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC) that can be connected to other network devices through a base station to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (Radio Frequency, RF for short) module, and the RF module is configured to communicate with the Internet wirelessly.

在本实施例中提供了一种运行于终端的网页预下载方法。图2是根据本申请实施例提供的一种网页预下载方法的流程图，如图2所示，该流程包括如下步骤：In this embodiment, a webpage pre-downloading method running on a terminal is provided. 2 is a flowchart of a method for pre-downloading a webpage according to an embodiment of the present application. As shown in FIG. 2, the process includes the following steps:

步骤S202，在终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，所述终端根据网页的内容属性确定需要预下载的网页；Step S202, in the case that the terminal currently connects to the webpage pre-downloading time point through the wireless fidelity WIFI connection network and/or the current time, the terminal determines the webpage that needs to be pre-downloaded according to the content attribute of the webpage;

步骤S204，所述终端下载所述确定的网页。Step S204, the terminal downloads the determined webpage.

通过该方案，能够在有WIFI和/或到达网页预下载时间点的情况下，基于网页的内容属性确定哪些网页要进行预下载。由于进行预下载的网页是根据网页的内容属性来确定的，可以实现更加有针对性的网页预下载，例如，可以有针对性地筛选涉及用户感兴趣或可能感兴趣的内容的网页进行预下载，也可以有针对性地针对目前内容热度高的网页进行预下载等等，在避免下载全部网页导致内存耗尽的情况下，保证用户在后续需要浏览网页时的浏览体验，可以解决相关技术中用户浏览网页需使用移动流量或打开缓慢的问题，提高了用户的体验。Through this scheme, it is possible to determine which web pages are to be pre-downloaded based on the content attributes of the web pages in the presence of WIFI and/or arrival web page pre-download time points. Since the pre-downloaded webpage is determined according to the content attribute of the webpage, a more targeted webpage pre-downloading can be implemented, for example, a webpage that is related to content that is of interest or may be of interest to the user can be targeted for pre-downloading. It can also be targeted for pre-downloading of web pages with high content content, etc., in the case of avoiding downloading all webpages and causing memory exhaustion, ensuring the browsing experience of the user when subsequently browsing the webpage, and solving related technologies Users browsing the web need to use mobile traffic or slow open issues to improve the user experience.

在步骤S202中，可以通过偏好度评估模型来确定选择哪个或哪些网页进行预下载更符合用户本身对于网页内容的喜好。例如，所述终端可以通过以下方式根据网页的内容属性确定需要预下载的网页：In step S202, it may be determined by the preference evaluation model to select which web page or web pages to pre-download more in line with the user's own preferences for web content. For example, the terminal may determine a webpage that needs to be pre-downloaded according to the content attribute of the webpage in the following manner:

所述终端解析当前网站主页，确定待评估的网页，作为一种实施方式，可以解析当前网站主页中所包含的全部链接或者当前网站主页中预定板块包含的全部链接，将这些链接对应的网页作为待评估的网页；The terminal parses the current website homepage and determines the webpage to be evaluated. As an implementation manner, all the links included in the current website homepage or all the links included in the predetermined section of the current website homepage may be parsed, and the webpage corresponding to the links is used as the webpage. The web page to be evaluated;

所述终端将待评估的网页中每个网页的内容属性输入偏好度评估模型，得到每个网页对应的偏好度评估值；The terminal inputs the content attribute of each webpage in the webpage to be evaluated into the preference evaluation model, and obtains a preference evaluation value corresponding to each webpage;

所述终端根据所述偏好度评估值对所述待评估的网页进行排序，选择排序靠前的N个网页作为所述需要预下载的网页，或者，所述终端选择所述待评估的网页中偏好度评估值高于预下载阈值的网页作为所述需要预下载的网页，其中，N为正整数。The terminal sorts the web pages to be evaluated according to the preference evaluation value, selects the N pages that are ranked first as the web pages that need to be pre-downloaded, or the terminal selects the webpage to be evaluated. The webpage with the preference evaluation value higher than the pre-download threshold is used as the webpage that needs to be pre-downloaded, where N is a positive integer.

在该实施例中，N的值可以根据所述终端的内存容量和/或平均上网时长确定。该N的值可以根据当前终端的内存容量和/或最新统计的用户的平均上网时长进行实时调整或者定期调整，从而保证预下载的网页量能够适应当前终端的硬件运行情况和/或用户对阅读量的需求。In this embodiment, the value of N may be determined according to the memory capacity of the terminal and/or the average time length of the Internet. The value of the N may be adjusted in real time or periodically according to the memory capacity of the current terminal and/or the average online time of the user of the latest statistics, thereby ensuring that the amount of pre-downloaded webpages can adapt to the hardware operation of the current terminal and/or the user reads The amount of demand.

根据网页的内容属性确定需要预下载的网页的过程所基于的模型可以是基于用户行为自动调节和学习的。作为一种实施方式，所述偏好度评估模型是使用多组训练样本通过机器学习训练出的，所述多组训练样本中的每组训练样本均包括：用户已浏览的网页的内容属性作为输入向量，以及基于所述已浏览的网页的停留时长计算得到的偏好度评估值作为输出标签。The model upon which the process of pre-downloading a web page is required based on the content attributes of the web page may be based on user behavior automatically adjusted and learned. As an implementation manner, the preference evaluation model is trained by using a plurality of sets of training samples, each of the plurality of sets of training samples includes: a content attribute of a web page that the user has browsed as an input. a vector, and a preference evaluation value calculated based on a dwell time of the browsed web page as an output tag.

所述偏好度评估模型可以为：y＝k1*x1+k2*x2+k3*x3+…+kn*xn，其中，所述偏好度评估模型的所述多组训练样本中的每组训练样本包括：用户已浏览的网页的内容属性在n个维度的属性值x1、x2、x3…xn作为输入向量，以及所述已浏览的网页对应的偏好度评估值y＝aT作为输出标签，其中，T为用户浏览所述已浏览的网页的停留时长，a为调节参数，用于调节所述y＝aT的计算结果在预定数值范围内，k1、k2、k3...kn为基于所述多组训练样本训练出来的所述内容属性在各个维度的权值，n为正整数。The preference evaluation model may be: y=k1*x1+k2*x2+k3*x3+...+kn*xn, wherein each of the plurality of sets of training samples of the preference evaluation model includes The content attribute of the web page that the user has browsed has the attribute values x1, x2, x3, . . . xn in n dimensions as an input vector, and the preference evaluation value y=aT corresponding to the browsed web page as an output label, where T For the user to browse the stay duration of the browsed webpage, a is an adjustment parameter, and the calculation result for adjusting the y=aT is within a predetermined numerical range, and k1, k2, k3...kn are based on the multiple groups The weight of the content attribute trained in the training sample in each dimension, n is a positive integer.

在对模型进行训练和学习时所基于的训练样本的输入向量，以及在基于网页的内容属性确定是否需要对该网页进行预下载的情况下，使用的所述网页的内容属性可以包括以下至少之一：所述网页中内容的类别、所述网页中内容的热度、所述网页中内容的时效性。The input vector of the training sample on which the model is trained and learned, and in the case where the web-based content attribute determines whether the web page needs to be pre-downloaded, the content attribute of the web page used may include at least the following A: a category of content in the webpage, a heat of content in the webpage, and a timeliness of content in the webpage.

总之，可以利用各种方式使得所述步骤S204可包括：将获取的网页的内容属性与用户喜好查阅的喜好属性一致的内容，作为需要预先现在的网页。In summary, the step S204 may be performed in various manners, and the content that matches the content attribute of the obtained web page with the favorite attribute of the user's favorite view may be used as a web page that needs to be present in advance.

作为一种实施方式，所述网页的内容属性包括所述网页中内容的类别的情况下，所述网页中内容的类别可以通过以下方式至少之一确定：In an embodiment, if the content attribute of the webpage includes a category of the content in the webpage, the category of the content in the webpage may be determined by at least one of the following manners:

根据所述网页的网址进行关键字匹配，从而确定所述网页中内容的类别；Performing keyword matching according to the web address of the webpage to determine the category of the content in the webpage;

对所述网页的内容中的预定成分进行分词处理，并对分词处理的结果进行类别分析，从而确定所述网页中内容的类别，例如，可以通过分词处理，确定内容所属的类别，如体育、财经、育儿等等。Performing word segmentation processing on predetermined components in the content of the webpage, and performing category analysis on the result of the word segmentation process, thereby determining a category of the content in the webpage, for example, determining, by using word segmentation, a category to which the content belongs, such as sports, Finance, child care, etc.

作为一种实施方式，所述网页的内容属性包括所述网页中内容的热度的情况下，所述网页中内容的热度可以通过以下方式确定：In an embodiment, if the content attribute of the webpage includes the popularity of the content in the webpage, the popularity of the content in the webpage may be determined by:

根据所述网页对应的链接在网站主页中的位置排名，确定所述网页中内容的热度，例如，位置排名靠前的链接对应的网页中所包含的内容的热度高。Determining the popularity of the content in the webpage according to the location ranking of the link corresponding to the webpage in the website homepage, for example, the content included in the webpage corresponding to the top-ranked link is hot.

所述热度可以用于网页被终端设备下载和/或显示的次数或频次来体现，被终端设备下载和/或显示的次数或频次越高，则热度越高。热度越高的网页内容，则越有可能是用户感兴趣的内容，即便在没有WiFi覆盖的情况下依然想要阅读的内容。作为一种实施方式，所述网页的内容属性包括所述网页中内容的时效性的情况下，所述网页中内容的时效性可以通过以下方式确定：The heat may be used to display the number or frequency of downloading and/or displaying of the webpage by the terminal device, and the higher the number or frequency of downloading and/or displaying by the terminal device, the higher the heat. The higher the content of the web page, the more likely it is that the user is interested in the content, even if there is no WiFi coverage. In an embodiment, when the content attribute of the webpage includes the timeliness of the content in the webpage, the timeliness of the content in the webpage may be determined by:

根据所述网页对应的发布时间和/或所述网页中内容的类别，确定所述网页中内容的时效性，例如，可以根据新闻或资讯本身内容的有效期来确定是否对该页面进行预下载，例如，雾霾预告或更新报告，如果预下载的时间点已超过雾霾预定时间点，则可以不必列入下载，或者雾霾时间点与当前下载时间点高度相关，可以决定将其列入下载等等，内容的有效期可以通过网页对应的发布时间和/或网页中内容的类别来确定。Determining the timeliness of the content in the webpage according to the publishing time corresponding to the webpage and/or the category of the content in the webpage, for example, determining whether to pre-download the webpage according to the validity period of the news or the information itself. For example, if the pre-download time point has exceeded the smog scheduled time point, you may not need to include the download, or the smog time point is highly correlated with the current download time point, you can decide to include it in the download. Etc., the validity period of the content can be determined by the publishing time corresponding to the webpage and/or the category of the content in the webpage.

为了提升用户浏览网页的反应速度，可以在所述终端下载所述确定的网页之后，在后台对所述下载的网页进行预加载，并将预加载后的网页保存到内存中。这样，后续响应于网页调用请求，所述终端可以从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。通过该方案，当用户点击这个网页对应的链接时预加载后的网页可以瞬间切换到前台显示出来。In order to improve the response speed of the user browsing the webpage, after the terminal downloads the determined webpage, the downloaded webpage may be preloaded in the background, and the preloaded webpage is saved in the memory. In this way, in response to the webpage invocation request, the terminal may retrieve the preloaded webpage corresponding to the webpage invocation request from the memory. Through this scheme, when the user clicks on the link corresponding to the webpage, the pre-loaded webpage can be instantly switched to the foreground display.

为了能够更好的适配用户的使用***均上网时段，例如可以是一天中用户上网频率最高的时段，所述网页预下载时间点可以设置为比用户的平均上网时段的开始时间点提前预设时长的时间点。In order to better adapt to the user's usage habits, the timing of pre-downloading can be automatically adjusted and decided according to the user behavior, thereby ensuring that the pre-downloading and/or pre-loading of the webpage is completed before the user browses. The average online time period of the user can be obtained by counting the time interval of the user, for example, the time period in which the user frequently accesses the Internet during the day, and the pre-downloading time point of the webpage can be set to be earlier than the starting time of the average online time of the user. The time of the time.

所述方法还包括：The method further includes:

根据检测的查阅请求，确定用户想要查阅的网页；该查阅请求可为基于用户操作生成的；Determining a webpage that the user wants to view according to the detected review request; the referral request may be generated based on a user operation;

若该网页已经下载到终端，则直接从终端存储预先下载网页的存储区域提取网页并显示。If the webpage has been downloaded to the terminal, the webpage is directly extracted from the storage area where the terminal stores the pre-downloaded webpage and displayed.

在一些实施例中，所述方法还可包括：In some embodiments, the method may further include:

若该网页未下载到终端，且终端当前连接到WiFi网络，则根据所述查阅请求从网络侧下载用户想要查阅的网页。If the webpage is not downloaded to the terminal, and the terminal is currently connected to the WiFi network, the webpage that the user wants to view is downloaded from the network side according to the referral request.

在还有一些实施例中，所述方法还可包括：In still other embodiments, the method can further include:

若网页未下载到终端，且终端当前未连接到WiFi网络，则根据终端已下载且未被查阅的网页进行内容推荐；If the webpage is not downloaded to the terminal, and the terminal is not currently connected to the WiFi network, the content recommendation is performed according to the webpage that the terminal has downloaded and is not consulted;

基于作用于所述内容推荐的查阅操作，显示所述内容推荐所对应的网页。A web page corresponding to the content recommendation is displayed based on a review operation acting on the content recommendation.

所述内容推荐包括：显示所述终端已下载且尚未查阅的网页的内容提要，或者，提示信息。The content recommendation includes: displaying a content summary of the webpage that the terminal has downloaded and has not yet consulted, or prompting information.

所述查阅操作包括但不限于，作用于所述内容推荐的点击操作或滑动操作等。以上仅是举例，具体实现的方式有多种，在此就不一一举例了。The review operation includes, but is not limited to, a click operation or a slide operation or the like acting on the content recommendation. The above is only an example, and there are various ways to implement the specific ones, and the examples are not exemplified herein.

在一些实施例中，所述方法还包括：In some embodiments, the method further includes:

根据下载的网页，生成可供匹配的第一索引；Generate a first index that can be matched according to the downloaded webpage;

若接收到所述查阅请求，将查阅请求指向网页的第二索引与所述第一索引匹配，若匹配成功，则说明终端本地有下载查阅请求所对应的网页。If the referral request is received, the second index of the reference requesting page is matched with the first index. If the matching is successful, the terminal has a webpage corresponding to the downloading request.

若未接收到所述查阅请求，则说明终端本地没有下载对应的网页。If the referral request is not received, the terminal does not download the corresponding webpage locally.

在还有一些实施例中，所述方法还包括：In still other embodiments, the method further includes:

根据终端本地显示已下载的网页的次数，生成和/或更新状态标记；其中，所述状态标记包括但不限于以下至少之一：指示已显示的第一显示标记；指示未显示的第二显示标记，指示已显示次数的第三显示标记；Generating and/or updating a status flag according to the number of times the terminal displays the downloaded web page locally; wherein the status flag includes but is not limited to at least one of: indicating a first display flag that has been displayed; indicating a second display that is not displayed a mark indicating a third display mark of the number of times displayed;

在对终端本地的网页进行推荐时，可以根据所述状态标记选择尚未显示本地存储的网页进行内容推荐。When recommending a webpage local to the terminal, the webpage that has not been displayed locally may be selected according to the status flag to perform content recommendation.

在一些实施例中，所述第一索引和所述状态标记可以对应存储在下载列表中，故若获得了查询请求可以将查询请求携带的第二索引，查阅所述下载列表。In some embodiments, the first index and the status flag may be correspondingly stored in a download list, so if a query request is obtained, the second index carried by the query request may be consulted.

所述第一索引和/或第二索引可为网页原始的统一资源地址等网页标识。在还有一些实施例中，还可以根据所述网页的内容属性，与用户的喜好属性进行匹配，生成匹配程度信息；然后将匹配程序信息加入到所述下载列表中，如此，后续在进行内容推荐时，可以根据匹配程序信息，优先推荐用户更加喜好的网页。The first index and/or the second index may be a webpage identifier such as a webpage original uniform resource address. In some embodiments, the matching attribute information may be generated according to the content attribute of the webpage to generate matching degree information; and then the matching program information is added to the downloading list, so that the content is subsequently performed. When recommending, you can give priority to recommending web pages that users prefer based on matching program information.

在还有一些实施例中，所述方法还包括以下至少之一：In still other embodiments, the method further includes at least one of the following:

若终端本地存储的网页达到预定数目，且当前连接到WiFi网络，删除已阅读的网页；If the webpage stored locally by the terminal reaches a predetermined number and is currently connected to the WiFi network, the read webpage is deleted;

或者，or,

利用新下载的网页内容覆盖已阅读或者已下载达到预定时长的网页。该已下载达到预定时长的网页包括：已下载达到预定时长且尚未阅读的内容。Use the newly downloaded web content to overwrite pages that have been read or have been downloaded for a predetermined length of time. The webpage that has been downloaded for a predetermined length of time includes content that has been downloaded for a predetermined length of time and has not yet been read.

通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到根据上述实施例的方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，或者网络设备等)执行本申请各个实施例所述的方法。Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware, but in many cases, the former is A better implementation. Based on such understanding, the technical solution of the present application, which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk, The optical disc includes a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present application.

在本实施例中还提供了一种运行于终端的网页预下载装置。该装置用于实现上述实施例及优选实施方式，已经进行过说明的不再赘述。如以下所使用的，术语“模块”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置较佳地以软件来实现，但是硬件，或者软件和硬件的组合的实现也是可能并被构想的。In the embodiment, a webpage pre-downloading device running on the terminal is also provided. The device is used to implement the above embodiments and preferred embodiments, and the description thereof has been omitted. As used below, the term "module" may implement a combination of software and/or hardware of a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware, is also possible and contemplated.

图3是根据本申请实施例提供的网页预下载装置的结构框图，如图3所示，该网页预下载装置可以应用于终端，所述装置可以包括：FIG. 3 is a structural block diagram of a webpage pre-downloading apparatus according to an embodiment of the present application. As shown in FIG. 3, the webpage pre-downloading apparatus may be applied to a terminal, and the apparatus may include:

网页预载决策模块32，配置为在所述终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，根据网页的内容属性确定需要预下载的网页；The webpage preloading decision module 32 is configured to determine, when the terminal currently connects to the webpage pre-downloading time point through the wireless fidelity WIFI connection network and/or the current time, determine the webpage that needs to be pre-downloaded according to the content attribute of the webpage;

网页下载模块34，耦合至所述网页预载决策模块32，配置为下载所述确定的网页。A web page download module 34, coupled to the web page preloading decision module 32, is configured to download the determined web page.

通过该方案，能够在有WIFI和/或到达网页预下载时间点的情况下，基于网页的内容属性确定哪些网页要进行预下载。由于进行预下载的网页是根据网页的内容属性来确定的，可以实现更加有针对性的网页预下载，在避免下载全部网页导致内存耗尽的情况下，保证用户在后续需要浏览网页时的浏览体验，可以解决相关技术中用户浏览网页需使用移动流量或打开缓慢的问题，提高了用户的体验。Through this scheme, it is possible to determine which web pages are to be pre-downloaded based on the content attributes of the web pages in the presence of WIFI and/or arrival web page pre-download time points. Since the pre-downloaded webpage is determined according to the content attribute of the webpage, a more targeted pre-downloading of the webpage can be realized, and in the case of avoiding the exhaustion of the memory by downloading all the webpages, the user is required to browse in the subsequent browsing of the webpage. The experience can solve the problem that the user needs to use mobile traffic or open slowly in browsing the webpage in related technologies, thereby improving the user experience.

图4是根据本申请实施例2的网页预下载装置的第一优选结构框图，如图4所示，该网页预下载装置还可以包括：待评估网页收集模块42，耦合至所述网页预载决策模块32，用于解析当前网站主页，确定待评估的网页并通知给所述网页预载决策模块32。4 is a block diagram of a first preferred structure of a webpage pre-downloading apparatus according to Embodiment 2 of the present application. As shown in FIG. 4, the webpage pre-downloading apparatus may further include: a webpage collecting module 42 to be evaluated, coupled to the webpage preloading The decision module 32 is configured to parse the current website homepage, determine the webpage to be evaluated, and notify the webpage preloading decision module 32.

所述网页预载决策模块32可以配置为：The webpage preloading decision module 32 can be configured to:

将所述待评估网页收集模块42通知的所述待评估的网页中每个网页的内容属性输入偏好度评估模型，得到每个网页对应的偏好度评估值；And inputting a content attribute of each webpage to be evaluated by the webpage to be evaluated by the to-be-evaluated webpage collecting module 42 into a preference evaluation model, to obtain a preference evaluation value corresponding to each webpage;

根据所述偏好度评估值对所述待评估的网页进行排序，将排序靠前的N个网页作为所述需要预下载的网页，或者，将所述待评估的网页中偏好度评估值高于预下载阈值的网页作为所述需要预下载的网页，其中，N为正整数。Sorting the webpages to be evaluated according to the preference evaluation value, using the N pages that are ranked first as the webpages that need to be pre-downloaded, or the preference evaluation value of the webpage to be evaluated is higher than The webpage pre-downloaded the threshold is used as the webpage that needs to be pre-downloaded, where N is a positive integer.

图5是根据本申请实施例2的网页预下载装置的第二优选结构框图。所述网页预载决策模块32根据网页的内容属性确定需要预下载的网页时所基于的模型可以是基于用户行为自动调节和学习的。作为一种实施方式，如图5所示，该网页预下载装置还可以包括：用户上网行为数据收集和学习模块52，用于使用多组训练样本通过机器学习训练出所述偏好度评估模型，所述多组训练样本中的每组训练样本均包括：用户已浏览的网页的内容属性作为输入向量，以及基于所述已浏览的网页的停留时长计算得到的偏好度评估值作为输出标签。FIG. 5 is a block diagram showing a second preferred structure of a webpage pre-downloading apparatus according to Embodiment 2 of the present application. The webpage preloading decision module 32 may determine that the model based on the webpage's content attributes based on the content attributes of the webpage may be automatically adjusted and learned based on the user behavior. As an embodiment, as shown in FIG. 5, the webpage pre-downloading apparatus may further include: a user online behavior data collection and learning module 52, configured to train the preference evaluation model by machine learning using a plurality of sets of training samples, Each of the plurality of sets of training samples includes: a content attribute of a web page that the user has browsed as an input vector, and a preference evaluation value calculated based on a dwell time of the browsed web page as an output tag.

所述用户上网行为数据收集和学习模块52在对模型进行训练和学习时所基于的训练样本的输入向量的情况下，以及在所述网页预载决策模块32基于网页的内容属性确定是否需要对该网页进行预下载的情况下，使用的所述网页的内容属性可以包括以下至少之一：所述网页中内容的类别、所述网页中内容的热度、所述网页中内容的时效性。The user online behavior data collection and learning module 52 determines, in the case of an input vector of the training sample upon which the model is trained and learned, and determines whether the web page preloading decision module 32 needs to be based on the content attribute of the web page. In the case that the webpage is pre-downloaded, the content attribute of the webpage used may include at least one of the following: a category of the content in the webpage, a heat of the content in the webpage, and a timeliness of the content in the webpage.

作为一种实施方式，所述网页的内容属性包括所述网页中内容的时效性的情况下，所述网页中内容的时效性可以通过以下方式确定：In an embodiment, when the content attribute of the webpage includes the timeliness of the content in the webpage, the timeliness of the content in the webpage may be determined by:

图6是根据本申请实施例2的网页预下载装置的第三优选结构框图，如图6所示，该网页预下载装置还可以包括：网页后台预加载子模块62，用于：FIG. 6 is a block diagram of a third preferred structure of a webpage pre-downloading apparatus according to Embodiment 2 of the present application. As shown in FIG. 6, the webpage pre-downloading apparatus 62 may further include: a webpage background preloading sub-module 62, configured to:

在后台对所述下载的网页进行预加载，并将预加载后的网页保存到内存中；The downloaded webpage is preloaded in the background, and the preloaded webpage is saved into the memory;

响应于网页调用请求，从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。The preloaded webpage corresponding to the webpage invocation request is retrieved from the memory in response to a webpage invocation request.

需要说明的是，上述各个模块是可以通过软件或硬件来实现的，对于后者，可以通过以下方式实现，但不限于此：上述模块均位于同一处理器中；或者，上述各个模块以任意组合的形式分别位于不同的处理器中。It should be noted that each of the above modules may be implemented by software or hardware. For the latter, the foregoing may be implemented by, but not limited to, the foregoing modules are all located in the same processor; or, the above modules are in any combination. The forms are located in different processors.

本申请的实施例还提供了一种存储介质，该存储介质中存储有计算机程序，其中，该计算机程序被设置为运行时执行上述任一项方法实施例中的步骤。Embodiments of the present application also provide a storage medium having stored therein a computer program, wherein the computer program is configured to execute the steps of any one of the method embodiments described above.

在本实施例中，上述存储介质可以被设置为存储用于执行以下步骤的计算机程序：In the present embodiment, the above storage medium may be arranged to store a computer program for performing the following steps:

S1，在终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，所述终端根据网页的内容属性确定需要预下载的网页；S1, in a case that the terminal currently connects to the webpage pre-downloading time point through the wireless fidelity WIFI connection network and/or the current time, the terminal determines the webpage that needs to be pre-downloaded according to the content attribute of the webpage;

S2，所述终端下载所述确定的网页。S2. The terminal downloads the determined webpage.

在一些实施例中，存储介质还被设置为存储用于执行以下步骤的计算机程序：In some embodiments, the storage medium is also arranged to store a computer program for performing the following steps:

S3，在所述终端下载所述确定的网页之后，在后台对所述下载的网页进行预加载，并将预加载后的网页保存到内存中；S3, after the terminal downloads the determined webpage, preloading the downloaded webpage in the background, and saving the preloaded webpage into a memory;

S4，响应于网页调用请求，所述终端可以从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。S4. In response to the webpage invocation request, the terminal may extract, from the memory, the preloaded webpage corresponding to the webpage invocation request.

在本实施例中，上述存储介质可以包括但不限于：U盘、只读存储器(Read-Only Memory，简称为ROM)、随机存取存储器(Random Access Memory，简称为RAM)、移动硬盘、磁碟或者光盘等各种可以存储计算机程序的介质。In this embodiment, the foregoing storage medium may include, but is not limited to, a USB flash drive, a Read-Only Memory (ROM), a Random Access Memory (RAM), a mobile hard disk, and a magnetic memory. A variety of media that can store computer programs, such as a disc or an optical disc.

本申请的实施例还提供了一种电子装置，包括存储器和处理器，该存储器中存储有计算机程序，该处理器被设置为运行计算机程序以执行上述任一项方法实施例中的步骤。Embodiments of the present application also provide an electronic device including a memory and a processor having a computer program stored therein, the processor being configured to execute a computer program to perform the steps of any of the above method embodiments.

在一些实施例中，上述电子装置还可以包括传输设备以及输入输出设备，其中，该传输设备和上述处理器连接，该输入输出设备和上述处理器连接。In some embodiments, the electronic device may further include a transmission device and an input and output device, wherein the transmission device is connected to the processor, and the input and output device is connected to the processor.

在本实施例中，上述处理器可以被设置为通过计算机程序执行以下步骤：In this embodiment, the above processor may be configured to perform the following steps by a computer program:

在本实施例中，上述处理器还可以被设置为通过计算机程序执行以下步骤：In this embodiment, the processor may be further configured to perform the following steps by using a computer program:

S4，响应于网页调用请求，所述终端可以从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。可选地，本实施例中的具体示例可以参考上述实施例及可选实施方式中所描述的示例，本实施例在此不再赘述。S4. In response to the webpage invocation request, the terminal may extract, from the memory, the preloaded webpage corresponding to the webpage invocation request. For example, the specific examples in this embodiment may refer to the examples described in the foregoing embodiments and the optional embodiments, and details are not described herein again.

本实施例提供了一种用于实现基于机器学习的网页预加载方法的***。This embodiment provides a system for implementing a machine learning based webpage preloading method.

图7是根据本申请实施例4的用于实现基于机器学习的网页预加载方法的***的结构框图。如图7所示，该***包括：用户上网行为数据收集和学习子模块70(其能够实现用户上网行为数据收集和学习模块52的功能)、网页链接地址收集子模块72(其能够实现待评估网页收集模块42的功能)、网页预加载决策子模块74(其能够实现网页预载决策模块32的功能)、网页下载子模块76(其能够实现网页下载模块34的功能)、网页后台预加载子模块78(其能够实现网页后台预加载子模块62的功能)。7 is a structural block diagram of a system for implementing a machine learning based webpage preloading method according to Embodiment 4 of the present application. As shown in FIG. 7, the system includes: a user online behavior data collection and learning sub-module 70 (which can implement the function of the user online behavior data collection and learning module 52), and a webpage link address collection sub-module 72 (which can implement the evaluation to be evaluated) The function of the webpage collecting module 42), the webpage preloading decision sub-module 74 (which can implement the function of the webpage preloading decision module 32), the webpage downloading sub-module 76 (which can implement the function of the webpage downloading module 34), and the webpage background preloading Sub-module 78 (which enables the functionality of web page background preload sub-module 62).

用户上网行为数据收集和学***常使用浏览器或其他新闻客户端的上网行为数据并进行学习和分析，如用户喜欢的网页对应哪一类型、用户常上网的时间段和时长，从而为网页预加载决策和时机提供反馈。The user online behavior data collection and learning sub-module 70 is mainly responsible for collecting the online behavior data of the terminal user or the other news client, and learning and analyzing, such as the type of the webpage that the user likes, the time period in which the user frequently accesses the Internet, and Duration, which provides feedback on web page preloading decisions and timing.

网页链接地址收集子模块72主要负责解析当前网站主页的链接地址，并根据用户上网行为时间分析反馈数据来实时更新需要待决策下载的链接地址。The webpage link address collection sub-module 72 is mainly responsible for parsing the link address of the current website homepage, and analyzing the feedback data according to the user's online behavior time to update the link address that needs to be downloaded in real time.

网页预加载决策子模块74主要负责通过训练的回归模型(回归模型可以用其他如神经网络模型替代)对待下载的网页链接地址进行综合打分排序，排序靠前的将通知网页下载模块对之进行提前下载和后台渲染。The webpage preloading decision sub-module 74 is mainly responsible for comprehensively sorting the downloaded webpage link addresses through the trained regression model (the regression model can be replaced by other such as the neural network model), and the prioritized webpage downloading module is notified in advance. Download and background rendering.

网页下载子模块76主要负责将需要预加载的网页从网站下载到本地。The webpage download sub-module 76 is primarily responsible for downloading webpages that need to be preloaded from the website to the local.

网页后台预加载子模块78主要负责对下载的网页在后台进行预加载操作(例如，可以包括对网页数据进行解码和解析、布局、GPU绘制等渲染工作，以及将要显示的内容准备好放到缓存中等等为能够正常显示网页所做的准备工作)，在用户点击这个网页对应的链接时可以瞬间切换到前台显示出来。The webpage background preloading submodule 78 is mainly responsible for preloading the downloaded webpage in the background (for example, may include rendering and parsing webpage data, layout, GPU rendering, etc., and preparing the content to be displayed in the cache) In the preparation work for displaying the webpage normally, the user can instantly switch to the foreground display when the user clicks on the link corresponding to the webpage.

该实施例提出了一种基于机器学习的网页预加载方法，它通过建立一种加权的回归模型来评估哪些网页需要在WIFI条件下提前加载和GPU绘制，该模型综合考虑用户对网页的喜好、网页热门度、以及网页的时效性等因素，相关的权值则通过机器学习来得到。This embodiment proposes a machine learning-based webpage preloading method, which establishes a weighted regression model to evaluate which webpages need to be loaded in advance and GPU rendering under WIFI conditions, and the model comprehensively considers the user's preference for the webpage, The popularity of the page, as well as the timeliness of the page, and related weights are obtained through machine learning.

该基于机器学习的网页预加载方法，至少包括两个阶段：一是离线学习阶段，即用户上网行为数据采集和学习分析阶段；二是在线决策阶段，即根据学习阶段提供的反馈数据来决策何时预加载哪些网页。以下分别对各个阶段的实现予以详细说明。The machine learning-based webpage preloading method includes at least two phases: one is an offline learning phase, that is, a user online behavior data collection and learning analysis phase; and the other is an online decision phase, that is, based on feedback data provided by the learning phase to decide whether Which pages are preloaded at the time. The implementation of each phase is described in detail below.

(一)离线学习阶段的实现(1) Implementation of the offline learning phase

离线学习阶段主要包括用户上网行为数据的采集、用户喜好网页综合评分的回归模型训练、用户上网时间段频率统计等操作，这部分工作主要由用户上网行为数据收集和学习子模块70来处理。The offline learning phase mainly includes the collection of user online behavior data, the regression model training of the user's favorite webpage comprehensive score, and the user online time period frequency statistics. This part of the work is mainly handled by the user online behavior data collection and learning sub-module 70.

用户上网行为数据的采集，主要是在浏览器或其他新闻类客户端进行埋点。通过埋点可实时地将用户上网行为的原始数据(数据结构如下表1)获取存储在本地数据库或发送回服务器进行分析处理。The collection of user online behavior data is mainly buried in the browser or other news client. Through the burying point, the original data of the user's online behavior (data structure is as shown in Table 1 below) can be obtained and stored in the local database or sent back to the server for analysis and processing.

用户标识User ID
网页地址website link
网页标题Page title
打开网页时间Open webpage time
网页停留时间Webpage time
链接地址在主页中的位置The location of the link address in the home page
....

表1采集用户上网行为原始数据的结构Table 1 Structure of collecting raw data of user online behavior

用户喜好网页综合评分的回归模型的训练主要需要构造：The training of the regression model of user-like webpage comprehensive scoring mainly needs to be constructed:

y＝k1*x1+k2*x2+k3*x3+…+kn*xn (式1)y=k1*x1+k2*x2+k3*x3+...+kn*xn (Equation 1)

其中输出y表示该用户对该网页的喜好程度，取值在0-1之间，值越大表示喜好程度越大。The output y indicates the degree of preference of the user to the webpage, and the value is between 0-1. The larger the value, the greater the preference.

输入(x1、x2、x3…xn)是一个多维输入向量，各分量表示该网页的类别(如体育、经济、旅游、热门等)，是个二值取值(0,1)，其中1表示该网页属于该类别，反之0则表示该网页不属于。The input (x1, x2, x3...xn) is a multi-dimensional input vector, each component representing the category of the web page (such as sports, economy, travel, popular, etc.), which is a binary value (0, 1), where 1 indicates the The web page belongs to the category, and the opposite 0 indicates that the web page does not belong.

k1、k2、k3...kn则是个输入各个维度(类别)的贡献权值，训练的过程就是通过对大量样本的学习得到这些最优的贡献权值，其中训练方法可采用常规的回归模型训练方法这里不再详细叙述，重点是说明各个训练样本(y:x1、x2、x3…xn)的构建。要构建各个训练样本，关键是收集的用户上网行为原始数据中每一条数据对应的网页进行分类。图8是根据本申请实施例5的输入向量的构造方法的流程图，如图8所示，具体实现步骤如下：K1, k2, k3...kn are the contribution weights of each input dimension (category). The training process is to obtain these optimal contribution weights by learning a large number of samples. The training method can adopt the conventional regression model. The training method is not described in detail here. The focus is on the construction of each training sample (y:x1, x2, x3...xn). To build each training sample, the key is to classify the web pages corresponding to each piece of data in the collected data of the user's online behavior. FIG. 8 is a flowchart of a method for constructing an input vector according to Embodiment 5 of the present application. As shown in FIG. 8, the specific implementation steps are as follows:

步骤S801：输入网址，和各类别关键字(如sport、finance、news等)进行匹配，如果匹配成功则将该网页划入该类并进入步骤S804，否则进入下一步骤S802处理；Step S801: input a web address, and match each category keyword (such as sport, finance, news, etc.), if the matching is successful, the webpage is classified into the class and proceeds to step S804, otherwise proceeds to the next step S802;

步骤S802：输入网页标题，进行分词处理；Step S802: input a webpage title to perform word segmentation processing;

步骤S803：通过分词建立该网页标题的词向量(Word Vector，将每个词表示为一个很长的向量。这个向量的维度是词表大小，其中绝大多数元素为0，只有一个维度的值为1，这个维度就代表了当前的词)，通过K近邻方法(k-Nearest Neighbor，简称为KNN，该方法的思路是：如果一个样本在特征空间中的k个最相似(即特征空间中最邻近)的样本中的大多数属于某一个类别，则该样本也属于这个类别)将该网页划入离该词向量距离最近的类别；需要说明的是，该K近邻分类方法还可以采用支持向量机等其他分类方法替代；Step S803: The word vector of the webpage title is established by word segmentation (Word Vector, each word is represented as a long vector. The dimension of the vector is a vocabulary size, wherein most of the elements are 0, and only one dimension value For 1, this dimension represents the current word), through the K-Nearest Neighbor method (KNN), the idea of this method is: if a sample is the most similar in the feature space, ie, in the feature space Most of the samples in the nearest neighbor belong to a certain category, and the sample also belongs to this category. The web page is classified into the category closest to the word vector; it should be noted that the K-nearest neighbor classification method can also support Alternatives to other classification methods such as vector machines;

步骤S804：输入链接地址在主页中的位置，如果为前几位就在热门类的维度上设置为1，否则为0；Step S804: input the location of the link address in the homepage, if it is the first few digits, set it to 1 in the dimension of the hot class, otherwise 0;

步骤S805：构造输入多维向量(x1、x2、x3…xn)，其中x1、x2、x3…xn-1,是各网页类别对应的值，如果是该类别就为1，否则为0；xn是热门维度的取值，如果是热门网页就为1，否则为0。Step S805: Constructing an input multi-dimensional vector (x1, x2, x3, ..., xn), where x1, x2, x3, ..., xn-1 are values corresponding to each web page category, and if it is the category, it is 1, otherwise 0; xn is The value of the hot dimension is 1 if it is a popular page, otherwise it is 0.

该方法能够根据用户的行为习惯选择性进行预加载，并提前进行GPU绘制，即节省流量和资源，对于打开网页的速度将是瞬间实现。本专利的涉及学习和决策时机都是根据用户行为自动调节，从而做到更加精准的网页预加载选择。The method can selectively preload according to the user's behavior habits, and perform GPU rendering in advance, that is, save traffic and resources, and the speed of opening the webpage will be realized instantaneously. The learning and decision-making timings of this patent are automatically adjusted according to user behavior, so as to achieve more accurate webpage preloading options.

通过上述步骤处理，就构建了各个学习样本的输入向量(x1、x2、x3…xn)。举个例子，假设该网页为体育类且是热门类的网页，则它的输入向量为：(1,0,0,….1)。其中第一个1表示为体育类，第二个1表示为热门类，其他0表示不是其他类别。Through the above steps, the input vectors (x1, x2, x3...xn) of the respective learning samples are constructed. For example, if the web page is a sports category and is a popular web page, its input vector is: (1, 0, 0, .... 1). The first one is represented as a sports class, the second one is represented as a popular class, and the other 0 is not a other category.

因为回归模型训练输入监督学习，还需要构造样本的输出标签y。这里我们将采集的原始数据结构中的网页停留时间T作为一个输出标签的影响因子，则输出标签y为：Since the regression model trains input supervised learning, it is also necessary to construct the output label y of the sample. Here we will take the webpage dwell time T in the original data structure collected as an influence factor of the output label, then the output label y is:

y＝aT(其中a是一个调节参数)。y=aT (where a is an adjustment parameter).

于是我们通过构建大量的样本数据(y:x1、x2、x3…xn)并使用常规的回归训练方法就得到该用户喜好网页综合评分的回归模型需要的权值k1、k2、k3...kn。So we construct the large amount of sample data (y:x1, x2, x3...xn) and use the conventional regression training method to get the weights k1, k2, k3...kn needed for the regression model of the user's favorite webpage comprehensive score. .

上述模型主要用于决策阶段判断预加载哪个网页，另外在离线阶段还需要通过对采集的用户上网时间的数据对用户上网时间段频率进行聚类统计，得到用户一天之中上网频率最高的时段，从而确定哪个时段进行网页的预加载时间(记为t1)，该时间一般设为用户上网频率最高时段的前半小时或一小时，这样可以获取主页中较新的链接网页。The above model is mainly used to determine which webpage is preloaded in the decision-making phase. In addition, in the offline phase, it is also necessary to perform clustering statistics on the frequency of the user's Internet access time by collecting the data of the online time of the user, and obtain the time period in which the user has the highest frequency of the Internet during the day. Therefore, it is determined which time period the preloading time (denoted as t1) of the webpage is performed, and the time is generally set to the first half hour or one hour of the highest frequency of the user's Internet access, so that the newer linked webpage in the homepage can be obtained.

(二)在线决策阶段(2) Online decision stage

在线决策阶段就是网页预加载决策子模块74通过离线阶段训练好的模型和其他参数来决定何时预加载哪些网页。图9是根据本申请实施例5的网页预加载决策子模块74的处理流程图。如图9所示，其主要处理步骤如下：The online decision phase is that the web page preloading decision sub-module 74 determines which web pages are preloaded by training the model and other parameters in the offline phase. FIG. 9 is a flowchart of processing of the webpage preloading decision sub-module 74 according to Embodiment 5 of the present application. As shown in Figure 9, the main processing steps are as follows:

步骤S901：判断当前上网环境是否处于WIFI，如果是则进入下一步骤，否则重复循环本步骤；需要说明的是，为了防止陷入死循环，也可以设置同时进行步骤S901和S902的判断，在判断其中任何一个符合的情况下，进入下一步骤S903；Step S901: determining whether the current Internet environment is in WIFI, if yes, proceeding to the next step, otherwise repeating this step; it should be noted that, in order to prevent falling into an infinite loop, it may also be set to simultaneously perform the judgments of steps S901 and S902, in judging If any one of them meets, proceed to the next step S903;

步骤S902：判断当前时间是否到了训练好的网页预加载时间t1，如果是则进入下一步骤S903，否则重复循环本步骤；Step S902: determining whether the current time has reached the trained webpage preloading time t1, and if yes, proceeding to the next step S903; otherwise, repeating this step;

步骤S903：启动网页链接地址收集子模块72，解析当前网站主页的链接地址和对应文本显示(对应链接的网页标题)，每个链接依次按照以下步骤S904和步骤S905处理；Step S903: The webpage link address collection sub-module 72 is started, and the link address of the current website homepage and the corresponding text display (corresponding to the linked webpage title) are parsed, and each link is sequentially processed according to the following steps S904 and S905;

步骤S904：同样按照上述训练中的步骤801至805构建该网页链接的输入向量(x1、x2、x3…xn)；Step S904: The input vector (x1, x2, x3, ... xn) of the webpage link is also constructed according to steps 801 to 805 in the above training;

步骤S905：将得到的输入向量(x1、x2、x3…xn)，代入按照训练好的回归模型公式(式1)计算该链接地址的用户喜好程度分数y；Step S905: Substituting the obtained input vector (x1, x2, x3, ..., xn) into the user preference degree y of the link address according to the trained regression model formula (Formula 1);

步骤S906：检查主页所有链接地址是否处理完毕，若是转入下一步骤，否则转入步骤S904继续处理；Step S906: Check whether all the link addresses of the homepage are processed, if the process proceeds to the next step, otherwise proceed to step S904 to continue processing;

步骤S907：将所有链接地址得到的喜好程度分数按照从大到小排序；Step S907: Sorting the preference scores obtained by all the link addresses according to the order of being large to small;

步骤S908：将排名最靠前的N个链接地址通知给网页下载子模块76进行下载，这里N的数目根据终端***的内存容量和用户平均上网时间共同决定，一般内存容量越大，N越大，同样用户上网时间越长，N也越大。Step S908: Notifying the top-ranking N link addresses to the webpage downloading sub-module 76 for downloading, where the number of N is determined according to the memory capacity of the terminal system and the average online time of the user. Generally, the larger the memory capacity, the larger the N The longer the same user goes online, the bigger the N is.

网页下载子模块76将收到的预加载网址通过终端网络模块在WIFI环境依次将这些网页数据从对应的服务器下载下来，并交给网页后台预加载子模块78对这些网页数据在后台进行解析、布局、GPU绘制等渲染工作。当用户点击这个网页对应的链接时可以瞬间切换到前台显示出来。The webpage downloading sub-module 76 downloads the pre-loaded webpages from the corresponding server in the WIFI environment through the terminal network module, and sends the webpage data to the webpage background pre-loading sub-module 78 to parse the webpage data in the background. Rendering work such as layout, GPU drawing, etc. When the user clicks on the link corresponding to this webpage, it can be instantly switched to the foreground display.

本实施例中的基于机器学习的网页预加载方法适用于所有终端上浏览器应用或者其他WEB应用及新闻客户端APP等。它通过建立一种加权的回归模型来评估哪些网页需要在WIFI条件下提前加载和GPU绘制，该模型综合考虑用户对网页的喜好、重要性、以及网页的时效性等因素，相关的权值则通过机器学习来得到。The machine learning-based webpage preloading method in this embodiment is applicable to all browser applications on the terminal or other WEB applications and news client APPs. It establishes a weighted regression model to evaluate which web pages need to be loaded in advance and GPU rendering under WIFI conditions. The model takes into account factors such as the user's preference for the web page, its importance, and the timeliness of the web page. Get it through machine learning.

该实施例中的方案与传统的网页预加载和普通的基于机器学习推荐系统不同之处在于：The solution in this embodiment differs from the traditional web page preloading and the general machine learning recommendation system in that:

1)根据用户的行为习惯选择性进行预加载，并提前进行GPU绘制，即节省流量和资源，对于打开网页的速度将是瞬间实现；1) Selectively preload according to the user's behavior habits, and perform GPU rendering in advance, that is, save traffic and resources, and the speed of opening the webpage will be instantaneous;

2)与普通基于机器学习推荐***比，该方法的涉及学习和决策时机都是根据用户行为自动调节，从而做到更加精准的选择。2) Compared with the general machine learning recommendation system, the learning and decision timing of the method is automatically adjusted according to the user behavior, so as to achieve more precise selection.

显然，本领域的技术人员应该明白，上述的本申请的各模块或各步骤可以用通用的计算装置来实现，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上，可选地，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，并且在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本申请不限制于任何特定的硬件和软件结合。Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, which can be concentrated on a single computing device or distributed in a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein. The steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps thereof are fabricated as a single integrated circuit module. Thus, the application is not limited to any particular combination of hardware and software.

以上所述仅为本申请的优选实施例而已，并不用于限制本申请，对于本领域的技术人员来说，本申请可以有各种更改和变化。凡在本申请的原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。The above description is only the preferred embodiment of the present application, and is not intended to limit the present application, and various changes and modifications may be made to the present application. Any modifications, equivalent substitutions, improvements, etc. made within the principles of this application are intended to be included within the scope of the present application.

Claims

一种网页预下载方法，包括：A webpage pre-download method includes:

在终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，In the case that the terminal currently connects to the network through the wireless fidelity WIFI and/or the current time reaches the pre-download time point of the webpage,

所述终端根据网页的内容属性确定需要预下载的网页；Determining, by the terminal, a webpage that needs to be pre-downloaded according to content attributes of the webpage

所述终端下载所述确定的网页。The terminal downloads the determined webpage.
根据权利要求1所述的方法，其中，所述终端根据网页的内容属性确定需要预下载的网页包括：The method of claim 1, wherein the determining, by the terminal, the webpage that needs to be pre-downloaded according to the content attribute of the webpage comprises:

所述终端解析当前网站主页，确定待评估的网页；The terminal parses the current website homepage to determine a webpage to be evaluated;

所述终端将待评估的网页中每个网页的内容属性输入偏好度评估模型，得到每个网页对应的偏好度评估值；The terminal inputs the content attribute of each webpage in the webpage to be evaluated into the preference evaluation model, and obtains a preference evaluation value corresponding to each webpage;

所述终端根据所述偏好度评估值对所述待评估的网页进行排序，选择排序靠前的N个网页作为所述需要预下载的网页，或者，所述终端选择所述待评估的网页中偏好度评估值高于预下载阈值的网页作为所述需要预下载的网页，其中，N为正整数。The terminal sorts the web pages to be evaluated according to the preference evaluation value, selects the N pages that are ranked first as the web pages that need to be pre-downloaded, or the terminal selects the webpage to be evaluated. The webpage with the preference evaluation value higher than the pre-download threshold is used as the webpage that needs to be pre-downloaded, where N is a positive integer.
根据权利要求2所述的方法，其中，N的值根据所述终端的内存容量和/或平均上网时长确定。The method of claim 2, wherein the value of N is determined based on a memory capacity of the terminal and/or an average length of time spent on the Internet.
根据权利要求2所述的方法，其中，所述偏好度评估模型是使用多组训练样本通过机器学习训练出的，所述多组训练样本中的每组训练样本均包括：用户已浏览的网页的内容属性作为输入向量，以及基于所述已浏览的网页的停留时长计算得到的偏好度评估值作为输出标签。The method according to claim 2, wherein the preference evaluation model is trained by machine learning using a plurality of sets of training samples, each of the plurality of sets of training samples comprising: a web page that the user has browsed The content attribute is used as an input vector, and a preference evaluation value calculated based on the dwell time of the browsed web page is used as an output tag.
根据权利要求4所述的方法，其中，所述偏好度评估模型为：y＝k ₁*x ₁+k ₂*x ₂+k ₃*x ₃+…+k _n*x _n，其中， The method according to claim 4, wherein said preference evaluation model is: y = k ₁ * x ₁ + k ₂ * x ₂ + k ₃ * x ₃ + ... + k _n * x _n , wherein

所述偏好度评估模型的所述多组训练样本中的每组训练样本包括：用户已浏览的网页的内容属性在n个维度的属性值x ₁、x ₂、x ₃…x _n作为输入向量，以及所述已浏览的网页对应的偏好度评估值y＝aT作为输出标签，其中，T为用户浏览所述已浏览的网页的停留时长，a为调节参数，用于调节所述y＝aT的计算结果在预定数值范围内，k ₁、k ₂、k ₃...k _n为基于所述多组训练样本训练出来的所述内容属性在各个维度的权值，n为正整数。 Each of the plurality of sets of training samples of the preference evaluation model includes: a content attribute of a web page that the user has browsed has attribute values x ₁ , x ₂ , x ₃ ... x _n in n dimensions as an input vector And the preference evaluation value y=aT corresponding to the browsed webpage is used as an output label, where T is a duration of the user browsing the browsed webpage, and a is an adjustment parameter, and is used to adjust the y=aT The calculation result is within a predetermined value range, and k ₁ , k ₂ , k ₃ ... k _n are weights of the content attributes trained in the plurality of sets of training samples in respective dimensions, and n is a positive integer.
根据权利要求1-5中任一项所述的方法，其中，所述网页的内容属性包括以下至少之一：所述网页中内容的类别、所述网页中内容的热度、所述网页中内容的时效性。The method according to any one of claims 1 to 5, wherein the content attribute of the webpage comprises at least one of: a category of content in the webpage, a heat of content in the webpage, content in the webpage Timeliness.
根据权利要求6所述的方法，其中，The method of claim 6 wherein

所述网页的内容属性包括所述网页中内容的类别的情况下，所述网页中内容的类别通过以下方式至少之一确定：Where the content attribute of the webpage includes a category of content in the webpage, the category of the content in the webpage is determined by at least one of the following manners:

根据所述网页的网址进行关键字匹配，从而确定所述网页中内容的类别；Performing keyword matching according to the web address of the webpage to determine the category of the content in the webpage;

对所述网页的内容中的预定成分进行分词处理，并对分词处理的结果进行类别分析，从而确定所述网页中内容的类别；Performing word segmentation processing on predetermined components in the content of the webpage, and performing category analysis on the result of the word segmentation processing, thereby determining a category of the content in the webpage;

所述网页的内容属性包括所述网页中内容的热度的情况下，所述网页中内容的热度通过以下方式确定：Where the content attribute of the webpage includes the popularity of the content in the webpage, the popularity of the content in the webpage is determined by:

根据所述网页对应的链接在网站主页中的位置排名，确定所述网页中内容的热度；Determining the popularity of the content in the webpage according to the location ranking of the link corresponding to the webpage in the website homepage;

所述网页的内容属性包括所述网页中内容的时效性的情况下，所述网页中内容的时效性通过以下方式确定：When the content attribute of the webpage includes the timeliness of the content in the webpage, the timeliness of the content in the webpage is determined by:

根据所述网页对应的发布时间和/或所述网页中内容的类别，确定所述网页中内容的时效性。Determining the timeliness of the content in the webpage according to the publishing time corresponding to the webpage and/or the category of the content in the webpage.
根据权利要求1-7中任一项所述的方法，其中，在所述终端下载所述确定的网页之后，还包括：The method according to any one of claims 1 to 7, wherein after the terminal downloads the determined webpage, the method further comprises:

所述终端在后台对所述下载的网页进行预加载，并将预加载后的网页保存到内存中；The terminal preloads the downloaded webpage in the background, and saves the preloaded webpage into a memory;

响应于网页调用请求，所述终端从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。In response to the webpage invocation request, the terminal retrieves the preloaded webpage corresponding to the webpage invocation request from the memory.
根据权利要求1-7中任一项所述的方法，其中，所述网页预下载时间点为比用户的平均上网时段的开始时间点提前预设时长的时间点。The method according to any one of claims 1 to 7, wherein the webpage pre-downloading time point is a time point that is longer than a preset time of a user's average online time period.
一种网页预下载装置，应用于终端，所述装置包括：A webpage pre-downloading device is applied to a terminal, the device comprising:

网页预载决策模块，用于在所述终端当前通过无线保真WIFI连接网络和/或当前时间到达网页预下载时间点的情况下，根据网页的内容属性确定需要预下载的网页；a webpage preloading decision module, configured to determine, according to a content attribute of the webpage, a webpage that needs to be pre-downloaded, if the terminal currently connects to the webpage pre-downloading time point through the wireless fidelity WIFI connection network and/or the current time;

网页下载模块，用于下载所述确定的网页。a webpage downloading module for downloading the determined webpage.
根据权利要求10所述的装置，其中，还包括：待评估网页收集模块，用于解析当前网站主页，确定待评估的网页并通知给所述网页预载决策模块；The device according to claim 10, further comprising: a webpage collection module to be evaluated, configured to parse the current website homepage, determine a webpage to be evaluated, and notify the webpage preloading decision module;

所述网页预载决策模块用于：The webpage preloading decision module is used to:

将所述待评估网页收集模块通知的所述待评估的网页中每个网页的内容属性输入偏好度评估模型，得到每个网页对应的偏好度评估值；And inputting a content attribute of each webpage in the webpage to be evaluated notified by the webpage collection module to be evaluated into a preference evaluation model, to obtain a preference evaluation value corresponding to each webpage;

根据所述偏好度评估值对所述待评估的网页进行排序，将排序靠前的N个网页作为所述需要预下载的网页，或者，将所述待评估的网页中偏好度评估值高于预下载阈值的网页作为所述需要预下载的网页，其中，N为正整数。Sorting the webpages to be evaluated according to the preference evaluation value, using the N pages that are ranked first as the webpages that need to be pre-downloaded, or the preference evaluation value of the webpage to be evaluated is higher than The webpage pre-downloaded the threshold is used as the webpage that needs to be pre-downloaded, where N is a positive integer.
根据权利要求11所述的装置，其中，N的值根据所述终端的内存容量和/或平均上网时长确定。The apparatus of claim 11, wherein the value of N is determined based on a memory capacity of the terminal and/or an average length of time spent on the Internet.
根据权利要求11所述的装置，其中，还包括：用户上网行为数据收集和学习模块，用于使用多组训练样本通过机器学习训练出所述偏好度评估模型，所述多组训练样本中的每组训练样本均包括：用户已浏览的网页的内容属性作为输入向量，以及基于所述已浏览的网页的停留时长计算得到的偏好度评估值作为输出标签。The apparatus according to claim 11, further comprising: a user online behavior data collection and learning module, configured to train the preference evaluation model by machine learning using a plurality of sets of training samples, wherein the plurality of sets of training samples Each set of training samples includes: a content attribute of a web page that the user has browsed as an input vector, and an evaluation value of the preference calculated based on the length of stay of the browsed web page as an output label.
根据权利要求13所述的装置，其中，所述偏好度评估模型为：y＝k ₁*x ₁+k ₂*x ₂+k ₃*x ₃+…+k _n*x _n，其中， The apparatus according to claim 13, wherein said preference evaluation model is: y = k ₁ * x ₁ + k ₂ * x ₂ + k ₃ * x ₃ + ... + k _n * x _n , wherein

所述偏好度评估模型的所述多组训练样本中的每组训练样本包括：用户已浏览的网页的内容属性在n个维度的属性值x ₁、x ₂、x ₃…x _n作为输入向量，以及所述已浏览的网页对应的偏好度评估值y＝aT作为输出标签，其中，T为用户浏览所述已浏览的网页的停留时长，a为调节参数，用于调节所述y＝aT的计算结果在预定数值范围内，k ₁、k ₂、k ₃...k _n为基于所述多组训练样本训练出来的所述内容属性在各个维度的权值，n为正整数。 Each of the plurality of sets of training samples of the preference evaluation model includes: a content attribute of a web page that the user has browsed has attribute values x ₁ , x ₂ , x ₃ ... x _n in n dimensions as an input vector And the preference evaluation value y=aT corresponding to the browsed webpage is used as an output label, where T is a duration of the user browsing the browsed webpage, and a is an adjustment parameter, and is used to adjust the y=aT The calculation result is within a predetermined value range, and k ₁ , k ₂ , k ₃ ... k _n are weights of the content attributes trained in the plurality of sets of training samples in respective dimensions, and n is a positive integer.
根据权利要求10-14中任一项所述的装置，其中，所述网页的内容属性包括以下至少之一：所述网页中内容的类别、所述网页中内容的热度、所述网页中内容的时效性。The device according to any one of claims 10-14, wherein the content attribute of the webpage comprises at least one of: a category of content in the webpage, a heat of content in the webpage, content in the webpage Timeliness.
根据权利要求15所述的装置，其中，The device according to claim 15, wherein

所述网页的内容属性包括所述网页中内容的类别的情况下，所述网页中内容的类别通过以下方式至少之一确定：Where the content attribute of the webpage includes a category of content in the webpage, the category of the content in the webpage is determined by at least one of the following manners:

根据所述网页的网址进行关键字匹配，从而确定所述网页中内容的类别；Performing keyword matching according to the web address of the webpage to determine the category of the content in the webpage;

对所述网页的内容中的预定成分进行分词处理，并对分词处理的结果进行类别分析，从而确定所述网页中内容的类别；Performing word segmentation processing on predetermined components in the content of the webpage, and performing category analysis on the result of the word segmentation processing, thereby determining a category of the content in the webpage;

所述网页的内容属性包括所述网页中内容的热度的情况下，所述网页中内容的热度通过以下方式确定：Where the content attribute of the webpage includes the popularity of the content in the webpage, the popularity of the content in the webpage is determined by:

根据所述网页对应的链接在网站主页中的位置排名，确定所述网页中内容的热度；Determining the popularity of the content in the webpage according to the location ranking of the link corresponding to the webpage in the website homepage;

所述网页的内容属性包括所述网页中内容的时效性的情况下，所述网页中内容的时效性通过以下方式确定：When the content attribute of the webpage includes the timeliness of the content in the webpage, the timeliness of the content in the webpage is determined by:

根据所述网页对应的发布时间和/或所述网页中内容的类别，确定所述网页中内容的时效性。Determining the timeliness of the content in the webpage according to the publishing time corresponding to the webpage and/or the category of the content in the webpage.
根据权利要求10-16中任一项所述的装置，其中，还包括网页后台预加载子模块，用于：The device according to any one of claims 10-16, further comprising a webpage background preloading submodule for:

在后台对所述下载的网页进行预加载，并将预加载后的网页保存到内存中；The downloaded webpage is preloaded in the background, and the preloaded webpage is saved into the memory;

响应于网页调用请求，从所述内存中取出对应于所述网页调用请求的所述预加载后的网页。The preloaded webpage corresponding to the webpage invocation request is retrieved from the memory in response to a webpage invocation request.
根据权利要求10-16中任一项所述的装置，其中，所述网页预下载时间点为比用户的平均上网时段的开始时间点提前预设时长的时间点。The apparatus according to any one of claims 10 to 16, wherein the webpage pre-downloading time point is a time point that is longer than a preset time of a user's average online time period.
一种存储介质，其中，所述存储介质中存储有计算机程序，其中，所述计算机程序被设置为运行时执行权利要求1至9任一项中所述的方法。A storage medium, wherein a computer program is stored in the storage medium, wherein the computer program is configured to perform the method of any one of claims 1 to 9 at runtime.
一种电子装置，包括存储器和处理器，其中，所述存储器中存储有计算机程序，所述处理器被设置为运行所述计算机程序以执行权利要求1至9任一项中所述的方法。An electronic device comprising a memory and a processor, wherein the memory stores a computer program, the processor being arranged to execute the computer program to perform the method of any one of claims 1 to 9.