An artificial intelligence data prediction system based on cloud computing

Abstract

The present invention provides a cloud computing-based artificial intelligence data prediction system, comprising a terminal acquisition module, a data storage module, a cloud processing module, and a terminal prediction module. The terminal acquisition module is used to collect user behavioral data; the data storage module is used to pre-process and store all user behavioral data; the cloud processing module obtains sample data from the data storage module and performs preliminary screening; and the terminal prediction module predicts the user's subsequent behavior based on the preliminary screening results and the behavioral data collected by the corresponding terminal acquisition module. This system considers the correlation between different behaviors and establishes a recording point network based on this correlation. The prediction process combines the sample data in the recording point network with the user's current behavioral data for calculation and processing, resulting in prediction results with high accuracy and low hysteresis.

Description

Technical Field

[0001] The present invention relates to the field of market research and analysis, and in particular to an artificial intelligence data prediction system based on cloud computing. Background Art

[0002] Internet services have significantly improved people's lives, especially recommendation algorithms, which can make targeted recommendations based on individual habits. However, existing recommendation algorithms primarily rely on statistics from a single user behavior to make recommendations. These recommendations remain largely consistent across time periods and require extensive behavioral change data to alter them, failing to meet user needs for timely access to the right services.

[0003] The foregoing discussion of the background art is intended only to facilitate an understanding of the present invention. This discussion does not acknowledge or admit that any of the material referred to is part of the common general knowledge.

[0004] Many recommendation systems have been developed. After extensive research and reference, we discovered existing recommendation prediction systems, such as the one disclosed in Publication No. CN109255079B. These systems generally include S1: receiving a personalized requirement Pk submitted by a cloud user Ui and selecting several most important attributes; S2: obtaining the cloud user's ratings for each cloud service J under personalized requirement Pk based on the user Ui's historical ratings in a rating database; and S3: recommending the highest-rated cloud service to the cloud user Ui. However, these systems fail to consider the temporal correlation between various user behaviors during recommendation processing, resulting in delayed predictions of user behavior and an inability to provide accurate cloud services in a timely manner. Summary of the Invention

[0005] The purpose of this invention is to address the existing deficiencies and propose an artificial intelligence data prediction system based on cloud computing.

[0006] The present invention adopts the following technical solutions:

[0007] An artificial intelligence data prediction system based on cloud computing, including a terminal acquisition module, a data storage module, a cloud processing module, and a terminal prediction module;

[0008] The terminal collection module is used to collect user behavior data, the data storage module is used to pre-process the behavior data of all users and then save it, the cloud processing module obtains sample data from the data storage module and performs preliminary screening, and the terminal prediction module predicts the user's subsequent behavior based on the preliminary screening results and the behavior data collected by the corresponding terminal collection module;

[0009] The data storage module includes a mapping unit, a storage unit, a weight calculation processor, and an index unit. The mapping unit is used to map behaviors in the collected data into recording points. The storage unit is used to save the associated network of recording points. The weight calculation processor is used to process weight calculation tasks between recording points. The index unit is used to retrieve the storage address of the recording point.

[0010] The cloud processing module obtains sample data from the associated network of the recording points and filters out the preparation points from the recording points;

[0011] The terminal prediction module calculates and processes the preparation points to obtain the final predicted behavior;

[0012] Furthermore, the storage unit creates a recording area for each recording point, the recording area including a preamble area, a basic information area, and a postamble area, the basic information area recording an index address of the recording area and number information of the recording point, the preamble area recording number information, cumulative weight value, and weight ratio of at least one previous recording point, and the postamble area recording number information, cumulative weight value, and weight ratio of at least one subsequent recording point;

[0013] Furthermore, the behavior data collected by the terminal collection module includes a set of execution points with a time sequence and a set of time intervals with a time sequence, each execution point in the set of execution points corresponds to a specific behavior, and each time interval in the set of time intervals represents the time difference between two adjacent execution points;

[0014] Furthermore, the weight calculation processor calculates the weight value Qu between the recording points corresponding to two adjacent execution points according to the following formula:

[0015]

[0016] Among them, q0 is the basic weight value, T0 is the standard time interval, t0 is the correction time interval, and t is the time interval between adjacent execution points;

[0017] The weight calculation processor accumulates each weight value to the cumulative weight value of the preceding area and the cumulative weight value of the following area of ​​the two corresponding recording areas;

[0018] The weight calculation processor recalculates the weight ratio R(i) based on the change of the accumulated weight value:

[0019]

[0020] Where Sum(i) represents the cumulative weight value of the recording point numbered i, where i is the number of the first recording point in the front region or the number of the last recording point in the back region;

[0021] Furthermore, the cloud processing module obtains at least two chain networks from the data storage module and calculates the reserve value P of each node in the chain network according to the following formula:

[0022]

[0023] Among them, n is the number of times the node appears in the chain network, R1(i) j is the first weight ratio of the node numbered i in the jth chain network, R2(i) is the second weight ratio of the node numbered i in the jth chain network, Lay j is the level of the node in the jth chain network, and L is the length of the chain network;

[0024] The cloud processing module takes the node whose reserve value exceeds the threshold as the reserve point

[0025] The calculation prediction unit calculates the prediction index Pca of each preparation point according to the following formula:

[0026]

[0027] Among them, m is the number of times the preparation point appears in the local storage unit, m k is the number of times the preparation point and the kth recording point collected in the detection window appear in the same time domain in the local storage unit, Nu is the sorting number of the preparation point in the local storage unit from high to low, and Nall is the number of all recording point types;

[0028] The terminal prediction module outputs the final prediction behavior of the preparation point with the largest prediction index as a result.

[0029] The beneficial effects achieved by the present invention are:

[0030] This system collects behavioral data from a large number of users and creates a recording point network. The recording point network records the relationship between each behavior. When making predictions for users, it first obtains the corresponding chain network data from the network based on the user's current behavior, and obtains preliminary points based on the relationship data of each recording point in the chain network. The preliminary points are then calculated and processed with the user's personal data to obtain the final prediction result. This prediction result integrates the data of all users and the personal data of the target user. When the user's current behavior changes, it can predict accurate results in a timely manner, reduce the hysteresis of the prediction result, and better recommend services.

[0031] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and illustration and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the overall structural framework of the present invention;

[0033] Figure 2 This is a schematic diagram of the data storage module of the present invention;

[0034] Figure 3 A schematic diagram of a process for predicting user behavior by the system of the present invention;

[0035] Figure 4 A schematic diagram of the process of obtaining a chain network for the present invention;

[0036] Figure 5 This is a schematic diagram of the terminal prediction module of the present invention. DETAILED DESCRIPTION

[0037] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. In addition, the drawings of the present invention are only simple schematic illustrations and are not depicted in actual size. It is stated in advance. The following embodiments will further explain the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0038] Example 1.

[0039] This embodiment provides an artificial intelligence data prediction system based on cloud computing, combined with Figure 1 , including terminal acquisition module, data storage module, cloud processing module, and terminal prediction module;

[0040] The terminal collection module is used to collect user behavior data, the data storage module is used to pre-process and save the behavior data of all users, the cloud processing module obtains sample data from the data storage module and performs preliminary screening, and the terminal prediction module predicts the user's subsequent behavior based on the preliminary screening results and the behavior data collected by the corresponding terminal collection module;

[0041] The data storage module includes a mapping unit, a storage unit, a weight calculation processor, and an index unit. The mapping unit is used to map behaviors in the collected data into recording points. The storage unit is used to save the associated network of recording points. The weight calculation processor is used to process weight calculation tasks between recording points. The index unit is used to retrieve the storage address of the recording point.

[0042] The cloud processing module obtains sample data from the associated network of the recording points and filters out the preparation points from the recording points;

[0043] The terminal prediction module calculates and processes the preparation points to obtain the final predicted behavior;

[0044] The storage unit creates a recording area for each recording point, the recording area including a preamble area, a basic information area, and a postamble area, the basic information area recording an index address of the recording area and number information of the recording point, the preamble area recording number information, cumulative weight value, and weight ratio of at least one previous recording point, and the postamble area recording number information, cumulative weight value, and weight ratio of at least one subsequent recording point;

[0045] The behavior data collected by the terminal collection module includes a set of execution points with a time sequence and a set of time intervals with a time sequence, each execution point in the set of execution points corresponds to a specific behavior, and each time interval in the set of time intervals represents the time difference between two adjacent execution points;

[0046] The weight calculation processor calculates the weight value Qu between the recording points corresponding to two adjacent execution points according to the following formula:

[0047]

[0048] Among them, q0 is the basic weight value, T0 is the standard time interval, t0 is the correction time interval, and t is the time interval between adjacent execution points;

[0049] The weight calculation processor accumulates each weight value to the cumulative weight value of the preceding area and the cumulative weight value of the following area of ​​the two corresponding recording areas;

[0050] The weight calculation processor recalculates the weight ratio R(i) based on the change of the accumulated weight value:

[0051]

[0052] Where Sum(i) represents the cumulative weight value of the recording point numbered i, where i is the number of the first recording point in the front region or the number of the last recording point in the back region;

[0053] The cloud processing module obtains at least two chain networks from the data storage module and calculates the reserve value P of each node in the chain network according to the following formula:

[0054]

[0055] Among them, n is the number of times the node appears in the chain network, R1(i) jis the first weight ratio of the node numbered i in the jth chain network, R2(i) is the second weight ratio of the node numbered i in the jth chain network, Lay j is the level of the node in the jth chain network, and L is the length of the chain network;

[0056] The cloud processing module takes the node whose reserve value exceeds the threshold as the reserve point

[0057] The calculation prediction unit calculates the prediction index Pca of each preparation point according to the following formula:

[0058]

[0059] Among them, m is the number of times the preparation point appears in the local storage unit, m k is the number of times the preparation point and the kth recording point collected in the detection window appear in the same time domain in the local storage unit, Nu is the sorting number of the preparation point in the local storage unit from high to low, and Nall is the number of all recording point types;

[0060] The terminal prediction module outputs the final prediction behavior of the preparation point with the largest prediction index as a result.

[0061] Example 2.

[0062] This embodiment includes all the contents of the first embodiment and provides an artificial intelligence data prediction system based on cloud computing, including a terminal acquisition module, a data storage module, a cloud processing module, and a terminal prediction module;

[0063] The terminal collection module is used to collect user behavior data, the data storage module is used to pre-process and save the behavior data of all users, the cloud processing module obtains sample data from the data storage module and performs preliminary screening, and the terminal prediction module predicts the user's subsequent behavior based on the preliminary screening results and the behavior data collected by the corresponding terminal collection module;

[0064] The terminal acquisition module and the terminal prediction module are integrated into a terminal device, the data storage module and the cloud processing module are integrated into a central server, and the system further includes a communication module for transmitting data between the terminal device and the central server;

[0065] The behavior data collected by the terminal collection module includes a set of execution points with a time sequence and a set of time intervals with a time sequence, each execution point in the set of execution points corresponds to a specific behavior, and each time interval in the set of time intervals represents the time difference between two adjacent execution points;

[0066] Combine Figure 2The data storage module includes a mapping unit and a storage unit, the mapping unit is used to map the execution point into a recording point, the recording point is used to represent a class of specific behaviors with high similarity, and the storage unit is used to store relevant information of the recording point;

[0067] The storage unit creates a recording area for each recording point, the recording area including a preamble area, a basic information area, and a postamble area, the basic information area recording an index address of the recording area and number information of the recording point, the preamble area recording number information, cumulative weight value, and weight ratio of at least one previous recording point, and the postamble area recording number information, cumulative weight value, and weight ratio of at least one subsequent recording point;

[0068] The preceding recording point and the following recording point are relative concepts. For two adjacent execution points in the behavior data, the recording point mapped to the preceding execution point is the preceding recording point of the recording point mapped to the following execution point, and the recording point mapped to the following execution point is the following recording point of the recording point mapped to the preceding execution point.

[0069] Combine Figure 2 The data storage module further includes a weight calculation processor, which calculates a weight value Qu between the recording points corresponding to two adjacent execution points according to the following formula:

[0070]

[0071] Among them, q0 is the basic weight value, T0 is the standard time interval, t0 is the correction time interval, and t is the time interval between adjacent execution points;

[0072] The weight calculation processor accumulates each weight value to the cumulative weight value of the preceding area and the cumulative weight value of the following area of ​​the two corresponding recording areas;

[0073] The weight calculation processor recalculates the weight ratio R(i) based on the change of the accumulated weight value:

[0074]

[0075] Where Sum(i) represents the cumulative weight value of the recording point numbered i, where i is the number of the first recording point in the front region or the number of the last recording point in the back region;

[0076] The weight ratios of the front region and the rear region are calculated separately;

[0077] Combine Figure 2, the storage unit further includes an index unit, in which a mapping relationship between the number of each recording point and the index address of the corresponding recording area is recorded, and the index unit can retrieve the corresponding recording area based on the number information of the recording point;

[0078] Combine Figure 3 The process of predicting user behavior by the system includes the following steps:

[0079] S1. The terminal collection module collects at least two behaviors of the user within the detection window and sends the corresponding execution points to the data storage module;

[0080] S2. After the data storage module maps the execution points to recording points, it obtains a chain network of length L based on each recording point and sends the chain network data to the cloud processing module;

[0081] S3. The cloud processing module obtains multiple preparation points based on chain network data analysis and sends the preparation points to the terminal prediction module;

[0082] S4. The terminal prediction module screens the prepared points based on local data to finally obtain a predicted point, where both the prepared points and the predicted point are recorded points;

[0083] Combine Figure 4 In step S2, the process of the data storage module acquiring the chain network includes the following steps:

[0084] S21, retrieve and obtain a recording area of ​​a recording point, which is called a target area;

[0085] S22, determining the numbers of three recording points with the highest weight ratios in the post-position area of ​​each target area, and these recording points are called target recording points;

[0086] S23, retrieving the recording areas of the target recording points, and resetting the target areas to these recording areas;

[0087] S24, repeat step S22 and step S23 for a total of (L-1) times;

[0088] S25. The initial recording points and the subsequently determined recording points constitute nodes on the chain network, and the nodes and the weight ratios between the nodes constitute chain network data;

[0089] It should be noted that the connection relationship between nodes on the chain network is a one-way connection, and there are two weight ratios between two nodes, called the first weight ratio and the second weight ratio. For example, if node A is connected to node B, the first weight ratio is the percentage of the cumulative weight value of the record point corresponding to node A connected to the record point corresponding to node B to the sum of the cumulative weight values ​​of the record point corresponding to node A connected to all record points, and the second weight ratio is the percentage of the cumulative weight value of the record point corresponding to node A connected to the record point corresponding to node B to the sum of the cumulative weight values ​​of all record points connected to the record point corresponding to node B;

[0090] The cloud processing module selects the preparation points according to the following process:

[0091] S31. Count the number of times each node appears in the chain network n;

[0092] S32. Calculate the reserve value P of each node according to the following formula:

[0093]

[0094] Among them, R1(i) j is the first weight ratio of the node numbered i in the jth chain network, R2(i) is the second weight ratio of the node numbered i in the jth chain network, Lay j is the level of the node in the jth chain network;

[0095] S33, selecting nodes whose reserve values ​​exceed a threshold as reserve points;

[0096] Combine Figure 5 The terminal prediction module includes a local storage unit and a calculation prediction unit. The local storage unit is used to store the user data collected by the terminal collection module. The calculation prediction unit calculates the prediction index Pca of each preparation point according to the following formula:

[0097]

[0098] Among them, m is the number of times the preparation point appears in the local storage unit, m k is the number of times the preparation point and the kth recording point collected in the detection window appear in the same time domain in the local storage unit, Nu is the sorting number of the preparation point in the local storage unit from high to low, and Nall is the number of all recording point types;

[0099] The local storage unit has the same storage method as the data storage module, except that the local storage unit only stores data of a single user and the original data. The same time domain means that the interval between the behaviors corresponding to the two recording points is within the time threshold;

[0100] The terminal prediction module outputs the final prediction behavior of the preparation point with the largest prediction index as a result;

[0101] The prediction results of this system can be used as reference information for recommendation algorithms in various APPs.

[0102] The contents disclosed above are only preferred feasible embodiments of the present invention and do not limit the scope of protection of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention description and drawings are included in the scope of protection of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. An artificial intelligence data prediction system based on cloud computing, characterized in that: It includes terminal acquisition module, data storage module, cloud processing module and terminal prediction module; The terminal collection module is used to collect user behavior data, the data storage module is used to pre-process and save the behavior data of all users, the cloud processing module obtains sample data from the data storage module and performs preliminary screening, and the terminal prediction module predicts the user's subsequent behavior based on the preliminary screening results and the behavior data collected by the corresponding terminal collection module; The data storage module includes a mapping unit, a storage unit, a weight calculation processor, and an index unit. The mapping unit is used to map behaviors in the collected data into recording points. The storage unit is used to save the associated network of recording points. The weight calculation processor is used to process weight calculation tasks between recording points. The index unit is used to retrieve the storage address of the recording point. The cloud processing module obtains sample data from the associated network of the recording points and filters out the preparation points from the recording points; The terminal prediction module calculates and processes the preparation points to obtain the final predicted behavior; The cloud processing module obtains at least two chain networks from the data storage module and calculates the reserve value of each node in the chain network according to the following formula: : ； Among them, n is the number of times the node appears in the chain network, is the first weight ratio of the node numbered i in the jth chain network, is the second weight ratio of the node numbered i in the jth chain network, is the level of the node in the jth chain network, and L is the length of the chain network; The cloud processing module takes the node whose reserve value exceeds the threshold as the reserve point The prediction unit calculates the prediction index of each preparation point according to the following formula: : ； Where m is the number of times the preparation point appears in the local storage unit, is the number of times the preparation point and the kth record point collected in the detection window appear in the same time domain in the local storage unit, It is the sorting sequence number of the preparation point in the local storage unit from high to low. The number of all record point types; The terminal prediction module outputs the final prediction behavior of the preparation point with the largest prediction index as a result.

2. The cloud computing-based artificial intelligence data prediction system according to claim 1, characterized in that: The storage unit creates a recording area for each recording point, and the recording area includes a front area, a basic information area and a back area. The basic information area records the index address of the recording area and the numbering information of the recording point. The front area records the numbering information, cumulative weight value and weight ratio of at least one previous recording point. The back area records the numbering information, cumulative weight value and weight ratio of at least one subsequent recording point.

3. The cloud computing-based artificial intelligence data prediction system according to claim 2, characterized in that: The behavior data collected by the terminal collection module includes a time-series execution point set and a time-series time interval set. Each execution point in the execution point set corresponds to a specific behavior, and each time interval in the time interval set represents the time difference between two adjacent execution points.

4. The cloud computing-based artificial intelligence data prediction system according to claim 3, characterized in that: The weight calculation processor calculates the weight value between the recording points corresponding to two adjacent execution points according to the following formula : ； in, is the basic weight value, is the standard time interval, To correct the time interval, is the time interval between adjacent execution points; The weight calculation processor accumulates each weight value to the cumulative weight value of the preceding area and the cumulative weight value of the following area of ​​the two corresponding recording areas; The weight calculation processor calculates the weight ratio based on the change of the accumulated weight value Recalculate: ； in, Indicates the cumulative weight value of the record point numbered i, where i is the number of the previous record point in the front area or the number of the next record point in the back area.

Images