CN111753145A - Mobile application use prediction method based on time sequence mode - Google Patents
Mobile application use prediction method based on time sequence mode Download PDFInfo
- Publication number
- CN111753145A CN111753145A CN202010523249.6A CN202010523249A CN111753145A CN 111753145 A CN111753145 A CN 111753145A CN 202010523249 A CN202010523249 A CN 202010523249A CN 111753145 A CN111753145 A CN 111753145A
- Authority
- CN
- China
- Prior art keywords
- mobile application
- user
- mobile
- time
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 239000013598 vector Substances 0.000 claims abstract description 28
- 230000007246 mechanism Effects 0.000 claims abstract description 14
- 238000003062 neural network model Methods 0.000 claims abstract description 11
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000005265 energy consumption Methods 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a mobile application use prediction method based on a time sequence mode, which comprises the steps of obtaining label data of mobile application, and constructing a bipartite graph of a user and the mobile application according to mobile application attribute information and a mobile application use sequence; then, modeling a relation between the user and the mobile application by using a graph neural network model to obtain a node vector representation of the mobile application; then respectively embedding the node vector of the mobile application and the position, time, usage amount and the like of the mobile application used by the user to obtain corresponding embedded vectors; then inputting the embedded vector into an RNN model based on an attention mechanism to obtain the use preference of the user; and finally, constructing a DNN model and outputting whether the user uses the label of the current mobile application at the next moment. The invention can more comprehensively utilize the use data of the smart phone, so that the user operation becomes simple and efficient, and can plan and optimize the battery energy consumption and the mobile application search time in advance.
Description
Technical Field
The invention relates to the field of mobile application data analysis, in particular to a mobile application use prediction method based on a time sequence mode.
Background
With the rapid development of wireless communication and mobile technology, smart phones have become important tools for people to communicate with each other in daily life. Existing research has shown that the number of APPs installed on a user's smartphone is between 10 and 90, and even over 90, with an average of about 50. This enormous number makes finding a particular APP on a smartphone not simple. The prediction of the usage of the mobile APP refers to predicting the APP which is most likely to be used next, so that the operation of the user becomes very simple and efficient when the user uses the smart phone, and the battery energy consumption and the application search time can be planned and optimized in advance. Due to more factors of all aspects involved in the prediction, the prediction method is complex, uncertain and high, and the prediction result has larger error. Currently, no effective prediction method appears.
Disclosure of Invention
In view of the above drawbacks, the present invention provides a time-series-mode-based mobile application usage prediction method with simple calculation and high prediction results.
The invention relates to a mobile application use prediction method based on a time sequence mode, which is characterized by comprising the following steps: the method comprises the following steps: acquiring related data of a mobile application, wherein the data comprises an application name, an application category and an application label; constructing a bipartite graph of the user and the mobile application according to the attribute information of the mobile application and the mobile application sequence; modeling by using a graph neural network model, and constructing a relation between the mobile user and the mobile application to obtain a node vector representation of the mobile application; respectively embedding the node vector of the mobile application and the position, time and usage amount of the mobile application used by a user; inputting the embedded vector into an RNN model based on an attention mechanism to obtain the use preference of a user; and constructing a DNN model and outputting whether the user uses the label of the current mobile application at the next moment.
Further, in the bipartite graph, when a user uses an APP, an edge exists between the user and the APP to represent the relationship between the user and the APP; when a certain APP is installed by two or more users at the same time, an edge exists between the several APPs, node vector representation of the mobile APP is constructed through a graph neural network model, and each node has own characteristics.
Further, a mobile application usage prediction method based on a time sequence mode is disclosed, wherein node vector representations of mobile APPs are constructed according to the order of usage of the mobile APPs.
Further, the mobile application use prediction method based on the time sequence mode is characterized in that the graph neural network model is based on an information propagation mechanism, and each node updates the node state of the node by exchanging information with each other until a certain stable value is reached.
Further, a mobile application based on time sequence mode uses a prediction method, and the graph neural network enables each node to be aware of other nodes on the graph by iteratively updating hidden states of all nodes.
Further, a mobile application use prediction method based on a time sequence mode is constructed, and an RNN model based on an attention mechanism is as follows: based on mobile APP usage data, a RNN model is used to learn timing characteristics in user usage habits.
Further, a mobile application based on time-series patterns uses a prediction method, and the attention mechanism means that output vectors are obtained by weighted summation of each element according to the importance degree of each element.
The invention has the beneficial effects that: the invention discloses a mobile application use prediction method based on a time sequence mode, which comprises the steps of constructing a bipartite graph of a user and a mobile application through mobile application attribute information and a mobile application use sequence; then, modeling the relationship between the user and the mobile application by using a graph neural network model, and acquiring the use preference of the user; and finally, constructing a DNN model and outputting whether the user uses the label of the current mobile application at the next moment. The invention can more comprehensively utilize the use data of the smart phone, so that the user operation becomes simple and efficient, and can plan and optimize the battery energy consumption and the mobile application search time in advance.
Drawings
Fig. 1 is a flowchart of a method for predicting usage of a mobile application based on a time sequence mode according to an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
As shown in fig. 1, a mobile application usage prediction method based on a time-series pattern includes the following steps:
s1, acquiring the relevant data of the mobile application from the application store, including: application name, application category, application label.
And S2, constructing a bipartite graph of the user and the mobile application according to the attribute information of the mobile application and the mobile application sequence.
And S3, modeling the relationship between the user and the mobile application by using the graph neural network model to obtain the node vector representation of the mobile application. And constructing a bipartite graph of the user and the mobile application according to the attribute information of the mobile application and the mobile application sequence. When a user uses an APP, an edge exists between the user and the APP, which represents the relationship between the user and the APP. When a certain APP is installed by two or more users at the same time, an edge exists between the several APPs, the APP use sequence is taken into consideration, and node vector representations of the mobile APPs are constructed through a graph neural network model, wherein each node has own characteristics. The graph neural network model is based on an information propagation mechanism, each node updates the node state of the node by exchanging information with each other until a certain stable value is reached, and the output of the GNN is calculated at each node according to the current node state. The graph neural network makes each node aware of other nodes on the graph by iteratively updating the hidden states of all nodes. And obtaining the embedded characteristic vector representation of the mobile APP through model learning, thereby obtaining the relation between the user and the mobile APP. The specific process formula is as follows:
xn＝fw(ln，lco[n]，xne[n]，lne[n]) (1)
on＝gw(xn，ln)
x(t+1)＝Fw(x(t)，l) (2)
xn(t+1)＝fw(ln，lco[n]，xne[n](t)，lne[n]) (3)
on(t)＝gw(xn(t)，ln)，n∈N
wherein the state of the node n is
Indicating that the output of the node is onAnd (4) showing. f. ofwFor the current transfer function, gwIs the current output function. FwAs a global transfer function, GwFor a global output function, respectively fwAnd gwIn the form of a superposition.
And S4, embedding the node vector of the mobile application and the position, time, usage amount and the like of the mobile application used by the user respectively to obtain corresponding embedded vectors.
And S5, inputting the embedded vector into an RNN model based on an attention mechanism to obtain the use preference of the user. And based on the mobile APP use data, learning the time sequence characteristics in the use habit of the user by using an RNN model, thereby realizing the use prediction of the mobile APP. Specifically, a node vector of the mobile application and a location, time, usage amount, and the like of the mobile application used by the user are embedded, respectively, to obtain corresponding embedded vectors. The mobile APP usage prediction problem is then modeled as a supervised classification problem, which is defined below.
Given set of users U ═ U1，u2，...，unH, mobile APP set P ═ { P }1，p2，...，pmUser's use position at time t
And length of use
Wherein n, m, k and s respectively represent the number of users, the number of mobile APPs, the number of positions where the users are located and the time length s of the users in the k positions. Output APP Embedded vectors for graph neural networks
As another input to the GRU. The formula of GRU is simplified as follows.
In the attention mechanism, the output vector is obtained by weighted summation of the elements according to the importance degrees, as shown in the formula.
Where the parameter i represents the time of day, j represents the jth element in the sequence, TxRepresenting the length of the sequence, f (-) representing the length of the pair of elements xjThe coding of (2). a isijCan be seen as a probability, reflecting the element hjTo CiThe importance of (d) can be expressed using softmax, as shown in the formula.
Where e isijIt is the matching degree between the element with code and other elements which is reflected, when the matching degree is higher, the influence of the element on the element is more, and aijThe larger the value of (c).
The time sequence characteristics of the use habit of the user using the mobile APP can be obtained through the RNN model based on the attention mechanism
And S6, constructing a DNN model, and outputting whether the label of the current mobile application is used by the user at the next moment. And inputting an output result of the RNN model based on the attention mechanism into the DNN model, learning the complex interaction between the user and the mobile application through the DNN model, and outputting whether the user uses the label of the current mobile application at the next moment. The specific process formula is as follows:
where U, I and T represent the user, mobile application and time embedded vectors. n denotes the number of hidden layers, Z1Phi in the layer1，W1And b1The weight matrix and the offset vector correspond to an activation function (ReLU or tanh function) representing the DNN layer usage.
Claims (7)
1. A mobile application use prediction method based on a time sequence mode is characterized in that: the method comprises the following steps:
acquiring related data of a mobile application, wherein the data comprises an application name, an application category and an application label;
constructing a bipartite graph of the user and the mobile application according to the attribute information of the mobile application and the mobile application sequence;
modeling by using a graph neural network model, and constructing a relation between the mobile user and the mobile application to obtain a node vector representation of the mobile application;
respectively embedding the node vector of the mobile application and the position, time and usage amount of the mobile application used by a user;
inputting the embedded vector into an RNN model based on an attention mechanism to obtain the use preference of a user;
and constructing a DNN model and outputting whether the user uses the label of the current mobile application at the next moment.
2. The method of claim 1, wherein the mobile application usage prediction method based on the time-series pattern comprises: in the bipartite graph, when a user uses an APP, an edge exists between the user and the APP to represent the relationship between the user and the APP; when a certain APP is installed by two or more users at the same time, an edge exists between the several APPs, node vector representation of the mobile APP is constructed through a graph neural network model, and each node has own characteristics.
3. The method of claim 2, wherein the mobile application usage prediction method based on the time-series pattern comprises: and constructing the node vector representation of the mobile APP according to the using sequence of the mobile APP.
4. The method of claim 1, wherein the mobile application usage prediction method based on the time-series pattern comprises: the graph neural network model is based on an information propagation mechanism, and each node updates the node state of the node by exchanging information with each other until a certain stable value is reached.
5. The method of claim 1, wherein the mobile application usage prediction method based on the time-series pattern comprises: the graph neural network makes each node aware of other nodes on the graph by iteratively updating the hidden states of all nodes.
6. The method of claim 1, wherein the mobile application usage prediction method based on the time-series pattern comprises: the constructed RNN model based on the attention mechanism is as follows: based on mobile APP usage data, a RNN model is used to learn timing characteristics in user usage habits.
7. The method of claim 1, wherein the mobile application usage prediction method based on the time-series pattern comprises: the attention mechanism means that the output vector is obtained by weighting and summing the elements according to the importance degrees of the elements.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010523249.6A CN111753145A (en) | 2020-06-10 | 2020-06-10 | Mobile application use prediction method based on time sequence mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010523249.6A CN111753145A (en) | 2020-06-10 | 2020-06-10 | Mobile application use prediction method based on time sequence mode |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111753145A true CN111753145A (en) | 2020-10-09 |
Family
ID=72675795
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010523249.6A Pending CN111753145A (en) | 2020-06-10 | 2020-06-10 | Mobile application use prediction method based on time sequence mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111753145A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114177747A (en) * | 2021-12-02 | 2022-03-15 | 昆岳互联环境技术(江苏)有限公司 | Flue gas desulfurization sulfur dioxide concentration prediction method based on machine learning algorithm |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117440A (en) * | 2015-08-11 | 2015-12-02 | 北京奇虎科技有限公司 | Method and apparatus for determining to-be-recommended application (APP) |
| CN107770365A (en) * | 2017-08-30 | 2018-03-06 | 天津大学 | A kind of prediction of multiple features APP streams and recommendation method based on Android |
| EP3322209A1 (en) * | 2016-11-11 | 2018-05-16 | Telia Company AB | A method and a mobile user terminal for service quality prediction |
| CN108536572A (en) * | 2018-04-16 | 2018-09-14 | 浙江大学 | Smart mobile phone App based on AppUsage2Vec models uses prediction technique |
| CN110083769A (en) * | 2019-04-29 | 2019-08-02 | 西北工业大学 | A kind of mobile application recommended method that user oriented preference develops |
| US20190340521A1 (en) * | 2016-12-27 | 2019-11-07 | Huawei Technologies Co., Ltd. | Intelligent Recommendation Method and Terminal |
| CN111259231A (en) * | 2018-12-03 | 2020-06-09 | 珠海格力电器股份有限公司 | Recommendation method and device for application program |
-
2020
- 2020-06-10 CN CN202010523249.6A patent/CN111753145A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117440A (en) * | 2015-08-11 | 2015-12-02 | 北京奇虎科技有限公司 | Method and apparatus for determining to-be-recommended application (APP) |
| EP3322209A1 (en) * | 2016-11-11 | 2018-05-16 | Telia Company AB | A method and a mobile user terminal for service quality prediction |
| US20190340521A1 (en) * | 2016-12-27 | 2019-11-07 | Huawei Technologies Co., Ltd. | Intelligent Recommendation Method and Terminal |
| CN107770365A (en) * | 2017-08-30 | 2018-03-06 | 天津大学 | A kind of prediction of multiple features APP streams and recommendation method based on Android |
| CN108536572A (en) * | 2018-04-16 | 2018-09-14 | 浙江大学 | Smart mobile phone App based on AppUsage2Vec models uses prediction technique |
| CN111259231A (en) * | 2018-12-03 | 2020-06-09 | 珠海格力电器股份有限公司 | Recommendation method and device for application program |
| CN110083769A (en) * | 2019-04-29 | 2019-08-02 | 西北工业大学 | A kind of mobile application recommended method that user oriented preference develops |
Non-Patent Citations (5)
| Title |
|---|
| RISTY M. ACERADO 等: "PredICT: A Mobile Application for Predicting the Students" Career using Naïve Bayes Algorithm", 《ACM》 * |
| 孙悦 等: "移动APP演化策略研究", 《移动APP演化策略研究》 * |
| 张栗粽等: "面向金融数据的神经网络时间序列预测模型", 《计算机应用研究》 * |
| 张艺璇 等: "移动APP演化模式分析与预测", 《计算机科学与探索》 * |
| 张艺璇等: "基于多级注意力机制网络的app流行度预测", 《计算机研究与发展》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114177747A (en) * | 2021-12-02 | 2022-03-15 | 昆岳互联环境技术(江苏)有限公司 | Flue gas desulfurization sulfur dioxide concentration prediction method based on machine learning algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Forecast methods for time series data: a survey | |
| CN110866190B (en) | Method and device for training neural network model for representing knowledge graph | |
| Wu et al. | Evolving RBF neural networks for rainfall prediction using hybrid particle swarm optimization and genetic algorithm | |
| Wang et al. | Evolutionary extreme learning machine ensembles with size control | |
| Tsai et al. | ACODF: a novel data clustering approach for data mining in large databases | |
| Ahmed et al. | A review on deep sequential models for forecasting time series data | |
| CN113190688B (en) | Complex network link prediction method and system based on logical reasoning and graph convolution | |
| CN113194493B (en) | Wireless network data missing attribute recovery method and device based on graph neural network | |
| Jin et al. | Deep‐learning prediction model with serial two‐level decomposition based on Bayesian optimization | |
| CN115329211B (en) | Personalized interest recommendation method based on self-supervision learning and graph neural network | |
| Sahoo et al. | Space–time forecasting of groundwater level using a hybrid soft computing model | |
| CN114862010A (en) | Flow determination method, device, equipment and medium based on space-time data | |
| Nie et al. | Digital twin for transportation big data: a reinforcement learning-based network traffic prediction approach | |
| Zhang et al. | Graph-based traffic forecasting via communication-efficient federated learning | |
| Ferreira et al. | Genetic algorithm for reservoir computing optimization | |
| CN111753145A (en) | Mobile application use prediction method based on time sequence mode | |
| WO2023015658A1 (en) | Point of interest recommendation method and system based on brain-inspired spatiotemporal perceptual representation | |
| Zhang et al. | A novel approach of tensor‐based data missing estimation for Internet of Vehicles | |
| CN115658971A (en) | Attention mechanism-based multi-layer heterogeneous network node importance degree evaluation method | |
| CN114445692A (en) | Image recognition model construction method and device, computer equipment and storage medium | |
| CN115034459A (en) | Pedestrian trajectory time sequence prediction method | |
| Thavasimani et al. | Deep learning techniques: A case study on comparative analysis of various optimizers to detect bots from CRESCI-2017 dataset | |
| CN111091290A (en) | Mobile phone assembly process knowledge framework model construction method based on transfer learning | |
| KR102561345B1 (en) | Apparatus and method for controlling hvac based on drqn | |
| KR102554278B1 (en) | Method for predicting structure of protein using twist-based structure updates |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20201009 |