User on-demand behavior analysis method based on big data processing

Abstract

The invention relates to the technical field of data analysis, in particular to a user on-demand behavior analysis method based on big data processing, and the method comprises the steps: 1, setting account verification for a target website, recording the use data of different accounts, dividing the use data of the same account, dividing the use data of the accounts into different use data intervals, and storing the use data of the accounts in different use data intervals; each use data interval corresponds to one time of use; 2, acquiring a movement speed check value, a route check value, an area check value, a palm check value and a pressure check value to form a check data set; 3, any two check data sets are extracted for data comparison, the check data sets are divided into a plurality of disjoint user sets on the basis of comparison results, and each user set corresponds to one distinguishing user; and 4, setting a browsing database for each user set, inputting the on-demand data in the use data interval corresponding to the check data set into the corresponding browsing database, and recommending on-demand content to the user based on the browsing database.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and particularly to a method for analyzing user on-demand behavior based on big data processing. Background Art

[0002] "On-demand behavior" generally refers to the behavior of users actively selecting to watch or listen to a certain content on a video or audio platform, especially in the way of on-demand selection, rather than the content played at a predetermined or fixed time (such as live broadcast). Compared with traditional radio and television, on-demand allows users to choose the viewing time and content according to their own needs, without being restricted by the fixed broadcast time. Users can freely choose the content they want to watch, reflecting high personalization and autonomy.

[0003] Video or audio platforms can analyze user preferences based on their on-demand behavior and recommend favorite content to them. The main object of analysis is the behavior of the entire account. However, generally, there is more than one user using the same account. Especially in a family scenario, there are often multiple shared users, which leads to insufficient personalization in the existing analysis and recommendation methods. There is an urgent need for a method for analyzing user on-demand behavior that can distinguish users based on usage data, so as to provide personalized on-demand recommendations for different users of the same account. Summary of the Invention

[0004] Aiming at the above-mentioned shortcomings of the existing technology, the present invention provides a method for analyzing user on-demand behavior based on big data processing, which can effectively solve the problem that the existing technology cannot provide personalized recommendations for on-demand accounts with multiple users.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: The present invention provides a method for analyzing user on-demand behavior based on big data processing, including the following steps: Step 1: Set account verification for the target website, record the usage data of different accounts, and divide the usage data of the same account: Obtain the opening time and closing time of the target website in the usage data, and analyze in combination with the usage operation time corresponding to different operations and the duration of the playing state, and divide the account usage data into different usage data intervals, and each usage data interval corresponds to one usage; Step 2: Conduct user verification at the starting time of the usage data interval. The user verification requires the user to click multiple interactive buttons and enter a verification number. During the user verification process: Obtain the moving path of the mouse pointer and the distance of the interactive buttons, and analyze the moving speed verification value and the route verification value in combination with the click time of each interactive button; Divide the outer surface of the mouse device into multiple contact areas, obtain the pressed areas in the contact areas, and analyze the position distribution of the pressed areas to obtain an area verification value and a palm verification value; Analyze and calculate the pressure verification value based on the pressing pressure of the keyboard keys corresponding to the verification numbers; Obtain the movement speed verification value, route verification value, area verification value, palm verification value, and pressure verification value to jointly form a verification data set; Step 3: Extract any two verification data sets for data comparison, and divide the verification data sets into multiple non-overlapping user sets based on the comparison results. Each user set corresponds to a different user; Step 4: Set up a browsing database for each user set, enter the on-demand data within the corresponding usage data interval of the verification data set into the corresponding browsing database, and recommend on-demand content to the user based on the browsing database.

[0006] Furthermore, the process of dividing the usage data interval is as follows: S1: Establish a time horizontal axis, mark the opening time and closing time of the target website on the time horizontal axis and record them as , where i is the serial number of the opening time. When i = 1, it represents the first opening. The time interval between adjacent opening times and closing times is recorded as the usage interval ; S2: When any two adjacent usage intervals satisfy , is the preset first time difference threshold, merge the two usage intervals into one usage interval, and enter S3; S3: Analyze a single usage interval, uniformly record any operation instruction within the usage interval as a usage operation, mark the corresponding time points of each usage operation on the time horizontal axis as usage operation times , calculate the time difference between any two adjacent usage operation times and record it as the operation time difference . When the target website is playing music or video, it is recorded as the playing state. When the operation time difference satisfies , is the preset second time difference threshold, where represents the duration of the playing state between adjacent usage operation times. Use the two adjacent usage operation times corresponding to the operation time difference as the interval demarcation points to split the usage interval into two usage intervals, and enter S4; S4: Divide the usage data of the same account within different time ranges into different usage intervals to obtain multiple usage data intervals.

[0007] Furthermore, the triggering conditions for user verification include: Condition 1: When the current moment is the opening moment and the time difference from the last closing moment of the website is greater than the first time difference threshold; Condition 2: When the current moment is the operation moment and the time difference from the last operation moment is greater than or equal to the second time difference threshold; Condition 3: When the current moment is the operation moment and the time difference from the end moment of the playback state of the target website is greater than or equal to the second time difference threshold.

[0008] Further, the calculation processes of the moving speed verification value and the route verification value are as follows: During the user verification process, the click order of each interactive button is sequentially recorded as , where f represents the click order of the interactive button, and the coordinate positions of each interactive button are respectively recorded as ; Calculate the straight-line distance between adjacent two interactive buttons and record it as the button distance, calculate the time difference between the click times of adjacent two interactive buttons to obtain the button time difference, calculate the moving speed judgment value by dividing the button distance by the button time difference, obtain the mouse pointer movement route between adjacent two interactive buttons and calculate the route length and record it as the route distance, and divide the route distance by the button distance to obtain the route judgment value; Calculate the average value of multiple moving speed judgment values to obtain the moving speed verification value, and calculate the average value of multiple route judgment values to obtain the route verification value.

[0009] Further, the calculation processes of the palm width reference value and the palm verification value are as follows: Divide the outer surface of the mouse device into multiple contact areas, draw each contact area in the same plane graph and divide it into a left pressing area, a middle pressing area and a right pressing area. The position distribution of each contact area in the same plane graph is consistent with its distribution on the outer surface of the mouse device, and set an anchor point in the middle pressing area; During the user verification process, record the contact area that is pressed as the pressing area, mark the pressing area in the plane graph, and conduct independent analysis on the pressing areas in each pressing area to obtain the finger pulp anchor point and the fingertip anchor point corresponding to each pressing area; Connect multiple finger pulp anchor points and fingertip anchor points to construct a closed polygon, calculate the area of the closed polygon and record it as the area verification value, and record the palm width reference value and the finger length reference value corresponding to each pressing area as , where u = 1, 2, 3, corresponding to the left pressing area, the middle pressing area and the right pressing area respectively, and substitute it into the formula for calculation to obtain the palm verification value, where represents the palm verification value, are all preset weight coefficients.

[0010] Further, the independent analysis process of the pressing area is as follows: Obtain the coordinates of the centers of each pressing area and calculate the distances between their center coordinates and the anchor points respectively. Denote the center of the pressing area closest to the anchor point as the fingertip anchor point of this pressing area. Calculate the coordinate distance between the fingertip anchor point and the anchor point to obtain the palm width reference value. Select the center of the pressing area farthest from the fingertip anchor point and denote it as the fingertip anchor point of this pressing area. Calculate the coordinate distance between the fingertip anchor point and the fingertip anchor point to obtain the finger length reference value.

[0011] Furthermore, the calculation process of the pressure verification value is as follows: A pressure sensor is provided under each key of the keyboard device to measure the pressure when the key is pressed. Obtain the peak pressure when each verification number corresponds to the keyboard key being pressed and denote it as the pressing pressure. Calculate the average value of the pressing pressures corresponding to each verification number to obtain the pressure verification value.

[0012] Furthermore, the specific process of comparing the verification data sets is as follows: Number the verification data sets in the generation order and denote them successively as , where n is the record serial number of the verification data set, n = 0, 1, 2,..., m, and m + 1 is the total number of the verification data sets. represents the first recorded verification data set; Denote the two verification data sets as the first comparison set and the second comparison set respectively. Obtain the movement speed verification value, route verification value, area verification value, palm verification value, and pressure verification value in the first comparison set and perform normalization to denote them as respectively. Obtain the movement speed verification value, route verification value, area verification value, palm verification value, and pressure verification value in the second comparison set and denote them as respectively. After performing normalization processing, substitute them into the formula for calculation to obtain the difference reference value, where: is the difference reference value; h = 1, 2, 3, 4, 5; is the preset weight coefficient; represents the maximum value in; represents the minimum value in; When the difference reference value is less than the preset difference threshold, generate a data set similarity signal and divide the two verification data sets into the same user set.

[0013] A computer device includes a memory and a processor. The memory stores a computer program. The processor, when executing the computer program, implements the steps of the above method.

[0014] A computer-readable storage medium has a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0015] The technical solution provided by the present invention has the following beneficial effects compared with the known prior art: 1. The present invention constructs a verification data set jointly composed of a movement speed verification value, a route verification value, an area verification value, a palm verification value, and a pressure verification value. This verification data set can fully reflect the user's usage habits and hand structure, and thus facilitate the comparison and differentiation of different users. Compared with the prior art, it can more accurately distinguish users. In particular, by introducing the data collected by external devices (mouse, keyboard), it overcomes the defects in the prior art that are limited to the operating system internal and have a single data source.

[0016] 2. The comparison and division process of the present invention is carried out immediately after the user's verification. Therefore, it can divide the user into the corresponding user set before the user's further on-demand operation, without affecting the on-demand recommendation in the subsequent process. And dividing users based on the comparison results can determine whether the user is one of the known users and recommend personalized on-demand content for them, thereby enhancing the user's on-demand experience. Compared with the prior art, it can further improve the application of differential recommendation for videos and music, enabling different users to obtain recommended on-demand content that meets their personal preferences even when sharing the same music and video accounts, greatly enhancing the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0018] Figure 1 It is the overall method flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0020] The following further describes the present invention with reference to the embodiments.

[0021] Refer to Figure 1 , a method for analyzing user on-demand behavior based on big data processing, at least including the following steps: Step 1: Analyze based on user usage data to identify and distinguish users, where: Denote the analyzed website as the target website (the target website can be a video website or a music website, and users can click on different interactive windows on the target website to on-demand music and videos that match their preferences). Set up account verification for the target website (specifically, correct account and password need to be entered for login verification). When the user passes the account verification, grant the user the usage permission and record the account usage data (the usage data includes but is not limited to a series of operation instructions and data such as browsing records, play records, website closing and opening, mouse cursor movement trajectory, mouse clicks, and keyboard inputs). Furthermore, based on the account usage data of the same account, divide the account usage data into different usage data intervals, and each usage data interval corresponds to a separate usage, where: S1: Establish a time horizontal axis, mark the opening time and closing time of the target website on the time horizontal axis and denote them as , where i is the serial number of the opening time, i = 1 represents the first opening, and the time interval between adjacent opening times and closing times is denoted as the usage interval ; S2: When any two adjacent usage intervals satisfy , is the preset first time difference threshold, merge the two usage intervals into one usage interval, and enter S3; S3: Analyze for a single usage interval, uniformly denote any operation instruction within the usage interval as a usage operation, mark the corresponding time points of each usage operation on the time horizontal axis as usage operation times , calculate the time difference between any two adjacent usage operation times and denote it as the operation time difference , when the target website is playing music or video, denote it as the playing state. When the operation time difference satisfies , is the preset second time difference threshold, where represents the duration of the playing state between adjacent usage operation times. Use the two adjacent usage operation times corresponding to the operation time difference as the interval demarcation points to split the usage interval into two usage intervals, and enter S4; S4: Divide the usage data of the same account within different time ranges into different usage intervals (each usage interval corresponds to a time range) to obtain multiple usage data intervals.

[0022] It should be noted that by dividing multiple usage data intervals, the usage data of the same account can be divided based on multiple usage records, and the usage records are determined by comparing the intervals between user operations. This enables the full utilization and differentiation of user usage data, facilitating horizontal comparison in subsequent processing and further facilitating the identification of whether the user has changed.

[0023] Step 2: Conduct user verification at the starting moment of the usage data interval, analyze the usage data during the user verification process, and distinguish the users corresponding to each usage. Specifically: The user verification is triggered when any of the following trigger conditions is met: Condition 1: When the current moment is an opening moment and the time difference from the most recent closing moment of the website is greater than the first time difference threshold; Condition 2: When the current moment is an operation moment and the time difference from the most recent operation moment is greater than or equal to the second time difference threshold; Condition 3: When the current moment is an operation moment and the time difference from the end moment of the playback state of the target website is greater than or equal to the second time difference threshold.

[0024] It should be noted that the trigger conditions for user verification require the website to be opened or an operation to be performed, and these conditions all require active interaction by the user. Therefore, the common prerequisite for the trigger conditions is that the user is interacting with the target website. Consequently, user verification will inevitably generate interaction data with the user, facilitating further differentiation and analysis in the future.

[0025] The specific process of user verification is as follows: Part1: Display multiple interaction buttons and multiple verification numbers on the target website page. When the user clicks all the interaction buttons (using a mouse device) and enters all the verification numbers (using a keyboard device), a verification passed instruction is generated.

[0026] Part2: During the user verification process, the click order of each interaction button is sequentially recorded as , where f represents the click order of the interaction button, and the coordinate positions of each interaction button are respectively recorded as (A plane rectangular coordinate system is constructed on the target website page to mark the coordinate positions of each interaction button); Calculate the straight-line distance between adjacent two interaction buttons as the button distance, calculate the time difference between the click times of adjacent two interaction buttons to obtain the button time difference, calculate the movement speed judgment value by dividing the button distance by the button time difference, obtain the mouse pointer movement route between adjacent two interaction buttons and calculate the route length as the route distance, and divide the route distance by the button distance to obtain the route judgment value; Calculate the average value of multiple movement speed judgment values to obtain the movement speed verification value, and calculate the average value of multiple route judgment values to obtain the route verification value (each group of adjacent interaction buttons corresponds to a movement speed judgment value and a route judgment value. Since there are multiple interaction buttons, there are multiple groups of adjacent interaction buttons); It should be noted that there are often differences in user habits when users drag and use the mouse device. For example, some users drive the mouse device by rotating the palm with the wrist as the fulcrum, and some users drive the mouse device by swinging the forearm with the elbow as the fulcrum. These two methods will cause differences in movement speed due to the different sizes of the muscle groups involved in the force, and there will also be differences in mouse control accuracy and manipulation speed among different users. These differences will lead to differences in the movement speed verification value and the route verification value.

[0027] Part3: Divide the outer surface of the mouse device into multiple contact areas, and each contact area is correspondingly provided with an independent pressure sensor for monitoring the pressing action of the user when using the mouse. Draw each contact area in the same plane graph and divide it into a left pressing area, a middle pressing area, and a right pressing area (corresponding to the left surface, the upper surface, and the right surface of the mouse device respectively). The position distribution of each contact area in the same plane graph is consistent with its distribution on the outer surface of the mouse device (that is, this plane graph is the horizontal expansion graph of all contact areas). Set an anchor point in the middle pressing area (the anchor point is located at the center of the area); During the user verification process, the contact area that is pressed is recorded as the pressing area and marked in the plane graph. Analyze each pressing area in the pressing area independently to obtain the finger pulp anchor point and the fingertip anchor point corresponding to each pressing area: Obtain the coordinates of the center of each pressing area and calculate the distance between its center coordinates and the anchor point respectively. Record the center of the pressing area closest to the anchor point as the finger pulp anchor point of this pressing area. Calculate the coordinate distance between the finger pulp anchor point and the anchor point to obtain the palm width reference value. Select the center of the pressing area farthest from the finger pulp anchor point as the fingertip anchor point of this pressing area. Calculate the coordinate distance between the finger pulp anchor point and the fingertip anchor point to obtain the finger length reference value; Connect multiple finger pulp anchor points and fingertip anchor points to construct a closed polygon. Calculate the area of the closed polygon and record it as the area verification value. Record the palm width reference value and the finger length reference value corresponding to each pressing area as , where u = 1, 2, 3, corresponding to the left pressing area, the middle pressing area, and the right pressing area respectively. Substitute it into the formula for calculation to obtain the palm verification value, where represents the palm verification value, are all preset weight coefficients; It should be noted that the area verification value is obtained based on the distribution of the finger pulp anchor points and the fingertip anchor points, and the distribution of the finger pulp anchor points and the fingertip anchor points depends on the distribution of the contact points between the palm and the mouse when the user uses the mouse device. Different users often have different contact point distributions due to differences in palm size and palm shape when using the same mouse device. In addition, the palm verification value indirectly reflects the length and interval of the main functional fingers when the user operates the mouse. Through the palm verification value, users with different grasping postures and palm differences can be distinguished.

[0028] Part4: A pressure sensor is provided under each key of the keyboard device to measure the pressure when the key is pressed, and obtain the peak pressure (i.e., the maximum pressure received during the pressing process) when the keyboard key corresponding to each verification number is pressed, which is recorded as the pressing pressure. Calculate the average value of the pressing pressures corresponding to each verification number to obtain the pressure verification value; It should be noted that when different users perform input operations using the same keyboard device, due to their different usual usage habits and different force magnitudes, they often exhibit different pressing forces, so different users can be distinguished based on this value.

[0029] Part5: Obtain the movement speed verification value, route verification value, area verification value, palm verification value, and pressure verification value to jointly form a verification data set, and each use (i.e., each use data interval) corresponds to a verification data set.

[0030] It should be noted that by constructing a verification data set jointly composed of the movement speed verification value, route verification value, area verification value, palm verification value, and pressure verification value, this verification data set can fully reflect the user's usage habits and hand structure, and thus facilitate comparing and distinguishing different users. Compared with the prior art, it can distinguish users more accurately. In particular, by introducing the data collected by external devices (mouse, keyboard), it overcomes the defect of being limited to the operating system internal and having a single data source in the prior art.

[0031] Step 3: Number the verification data sets in the generation order and record them successively as , where n is the record serial number of the verification data set, n = 0, 1, 2,..., m, and m + 1 is the total number of verification data sets, represents the first recorded verification data set (i.e., the data set recorded when the account logs in for the first time); Extract any two verification data sets for data comparison, and divide the verification data sets into multiple non-intersecting user sets based on the comparison results. Each user set corresponds to a distinguishable user. The comparison process is as follows: Denote the two verification data sets as the first comparison set and the second comparison set respectively. Obtain the moving speed verification value, route verification value, area verification value, palm verification value, and pressure verification value in the first comparison set, and perform normalization, which are respectively denoted as ; obtain the moving speed verification value, route verification value, area verification value, palm verification value, and pressure verification value in the second comparison set, which are respectively denoted as ; after performing normalization processing, substitute them into the formula for calculation to obtain the difference reference value, where: is the difference reference value; h = 1, 2, 3, 4, 5; is the preset weight coefficient; represents the maximum value in; represents the minimum value in; When the difference reference value is less than the preset difference threshold, generate a data set similarity signal, and divide the two verification data sets into the same user set.

[0032] It should be noted that the comparison and division process is carried out immediately after the user performs user verification. Therefore, it can divide the user into the corresponding user set before the user performs further on-demand operations, without affecting the on-demand recommendations in the subsequent process. And dividing users based on the comparison results can determine whether the user is one of the known users and recommend personalized on-demand content for them, thereby improving the user's on-demand experience. Compared with the existing technology, it can further improve the application of video and music differential recommendations, enabling different users to also obtain recommended on-demand content that meets their personal preferences when sharing the same music and video accounts, greatly improving the user experience.

[0033] Step Four: Set up a browsing database for each user set, enter the on-demand data within the corresponding data interval of the verification data set into the corresponding browsing database, so as to form the usage databases corresponding to different users. Based on the convolutional neural network (CNN), extract the image and audio features in the on-demand data, and combine with the use of natural language processing (NLP) technology to extract the text features (such as titles, descriptions, comments, etc.) in the on-demand data. Build an on-demand recommendation model based on multiple features and recommend videos that the user may be interested in.

[0034] It should be noted that deep learning algorithms (especially convolutional neural network CNN and recurrent neural network RNN) are existing technologies and have been widely used in short video recommendations, especially in content feature extraction, sequence modeling, and personalized recommendations. Therefore, no more details will be elaborated here.

[0035] A computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps in the above method are implemented.

[0036] A computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, the steps in the above method are implemented.

[0037] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing user on-demand behavior based on big data processing, characterized in that: The following steps are involved: Step 1: Set up account verification for the target website, record usage data of different accounts, and divide usage data of the same account: Obtain the opening and closing times of the target website in the usage data, analyze the usage operation times corresponding to different operations and the duration of the playback state, and divide the account usage data into different usage data intervals, each of which corresponds to one usage; Step 2: Perform user verification at the starting point of the interval using the data. User verification requires the user to click multiple interactive buttons and enter a verification number. During the user verification process: Get the movement path of the mouse pointer and the distance of the interactive buttons, and analyze the click time of each interactive button to get the movement speed verification value and route verification value; Divide the outer surface of the mouse device into multiple contact areas, obtain the pressing area in the contact area, and obtain the area verification value and the palm verification value based on the position distribution analysis of the pressing area; The pressure verification value is obtained based on the analysis and calculation of the pressing pressure of the keyboard key corresponding to the verification number; Obtaining a movement speed verification value, a route verification value, an area verification value, a palm verification value, and a pressure verification value to form a verification data set; Step 3: extract any two verification data sets for data comparison, and divide the verification data sets into multiple disjoint user sets based on the comparison results, each user set corresponds to a distinct user; Step 4: Set up a browsing database for each user set, enter the on-demand data within the usage data interval corresponding to the verification data set into the corresponding browsing database, and recommend on-demand content to the user based on the browsing database.

2. According to the method for analyzing user on-demand behavior based on big data processing according to claim 1, it is characterized in that: The process of dividing the data interval is as follows: S1: Establish a time axis, mark the opening and closing time of the target website on the time axis and record them as , i is the serial number of the opening time, i=1 means the first opening, and the time interval between the adjacent opening and closing times is recorded as the usage interval ; S2: When any two adjacent use intervals Satisfaction between hour, is a preset first time difference threshold, the two usage intervals are merged into one usage interval, and the process goes to S3; S3: Analyze a single usage interval, record any operation instructions within the usage interval as a usage operation, and mark the time point corresponding to each usage operation on the horizontal axis of time as the usage operation time. , calculate the time difference between any two adjacent operation times and record it as the operation time difference , when the target website plays music or video, it is recorded as the playing state, and when the operation time difference meets hour, is the preset second time difference threshold, where Indicates the duration of the playback state between adjacent use operation moments, and uses the two adjacent use operation moments corresponding to the operation time difference as the interval demarcation point to split the use interval into two use intervals, and enters S4; S4: Divide the usage data of the same account within different time ranges into different usage intervals to obtain multiple usage data intervals.

3. According to the method for analyzing user on-demand behavior based on big data processing in claim 1, it is characterized in that: The trigger conditions for user verification include: Condition 1: When the current time is the opening time and the time difference from the last closing time of the website is greater than the first time difference threshold; Condition 2: When the current time is the operation time and the time difference from the most recent operation time is greater than or equal to the second time difference threshold; Condition three: when the current time is the operation time and the time difference from the end time of the playback state of the target website is greater than or equal to the second time difference threshold.

4. The method for analyzing user on-demand behavior based on big data processing according to claim 3 is characterized in that: The calculation process of the speed check value and route check value is as follows: During the user verification process, the click order of each interactive button is recorded in turn. , where f represents the click order of the interactive buttons, and the coordinate positions of each interactive button are recorded as ; Calculate the straight-line distance between two adjacent interactive buttons as the button distance, calculate the time difference between the click times of two adjacent interactive buttons to obtain the button time difference, calculate the button distance divided by the button time difference to obtain the speed judgment value, obtain the mouse pointer movement route between two adjacent interactive buttons and calculate the route length as the route distance, and divide the route distance by the button distance to obtain the route judgment value; The average of the plurality of movement speed judgment values ​​is calculated to obtain the movement speed verification value, and the average of the plurality of route judgment values ​​is calculated to obtain the route verification value.

5. The method for analyzing user on-demand behavior based on big data processing according to claim 4 is characterized in that: The calculation process of palm width reference value and palm verification value is as follows: The outer surface of the mouse device is divided into multiple contact areas, and each contact area is drawn in the same plane and divided into a left pressing area, a middle pressing area, and a right pressing area. The position distribution of each contact area in the same plane is consistent with its distribution on the outer surface of the mouse device, and an anchor point is set in the middle pressing area; During the user verification process, the pressed contact area is recorded as the pressed area, the pressed area is marked in the plane map, and the pressed area in each pressed area is analyzed independently to obtain the fingertip anchor point and fingertip anchor point corresponding to each pressed area; Connect multiple fingertip anchor points and fingertip anchor points to construct a closed polygon, calculate the area of ​​the closed polygon and record it as the area verification value, and record the palm width reference value and finger length reference value corresponding to each pressing area as , where u=1, 2, 3, corresponding to the left pressing area, the middle pressing area, and the right pressing area, respectively. Substitute into the formula The palm verification value is calculated in Indicates the palm check value, These are all preset weight coefficients.

6. The method for analyzing user on-demand behavior based on big data processing according to claim 5 is characterized in that: The independent analysis process of the pressed area is as follows: Get the coordinates of the center of each pressing area and calculate the distance between its center coordinates and the anchor point respectively; record the center of the pressing area closest to the anchor point as the fingertip anchor point of the pressing area; calculate the coordinate distance between the fingertip anchor point and the anchor point to obtain the palm width reference value; select the center of the pressing area farthest from the fingertip anchor point as the fingertip anchor point of the pressing area; calculate the coordinate distance between the fingertip anchor point and the fingertip anchor point to obtain the finger length reference value.

7. The method for analyzing user on-demand behavior based on big data processing according to claim 1 is characterized in that: The calculation process of the pressure check value is as follows: A pressure sensor is provided under each key of the keyboard device to measure the pressure when the key is pressed. The peak pressure when each verification digit corresponds to the keyboard key is obtained as the pressing pressure, and the average pressing pressure corresponding to each verification digit is calculated to obtain the pressure verification value.

8. The method for analyzing user on-demand behavior based on big data processing according to claim 1 is characterized in that: The comparison process of the verification data set is as follows: The check data sets are labeled in the order of generation and recorded as , where n is the record number of the verification data set, n=0,1,2,…,m, and m+1 is the total number of verification data sets. Represents the first recorded collation data set; The two verification data sets are respectively recorded as the first comparison set and the second comparison set, and the speed verification value, route verification value, area verification value, palm verification value and pressure verification value in the first comparison set are obtained and normalized and recorded as , obtain the speed check value, route check value, area check value, palm check value and pressure check value in the second comparison concentration and record them as , after normalization, enter the formula The difference reference value is calculated in, where: is the difference reference value; h=1,2,3,4,5； is the preset weight coefficient; express The maximum value in ; express The minimum value in ; When the difference reference value is less than a preset difference threshold, a data set similarity signal is generated, and the two verification data sets are divided into the same user set.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images