A time interval-based session information cutting method

Abstract

The application discloses a conversation information cutting method based on time interval and belongs to the technical field of data analysis. The cutting method comprises the following steps: acquiring a behavior sequence of a target user, wherein the behavior sequence comprises at least one behavior triggered by the user based on a client and arranged in time sequence; a cutting rule comprises a preset time interval when the behavior sequence is cut, and the preset time interval is a maximum time interval between two adjacent behaviors; and the behavior sequence is cut according to the target cutting rule to obtain at least one conversation. The application can easily convert continuous behaviors of a user into a conversation based on time interval, and the integrity and time requirement of the conversation are considered, so that the analysis requirement of conversation information is better met.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a method for segmenting session information based on time intervals. Background Technology

[0002] In user-driven client-side actions, segmenting a single action targeting a specific goal into a coherent whole based on chronological order can be termed session information segmentation. Currently, the most commonly used session information segmentation methods are interval-based segmentation and start / end-event segmentation.

[0003] Interval-based segmentation: This method involves segmenting user behavior data according to a defined time interval to obtain the segmented session information. The interval is usually defined based on operational experience or user habits. This method has low computational cost, high performance, and simple technical implementation, but its drawback is poor accuracy in session information segmentation. It often divides a single user interaction into several sessions or groups multiple user interactions into a single session.

[0004] The start-end event segmentation method offers a significant improvement in accuracy compared to the interval-time segmentation method. It defines start and end events and segments a series of events that meet the start and end criteria, along with their intermediate processes, into a single session segment. This method greatly improves the preservation of session information integrity. However, its drawback is that it cannot fully predict user behavior patterns and cannot guarantee that a user will definitely generate start and end events. For example, if creating an order is defined as the start event and payment completion as the end event, if the user abandons the payment, the segmentation conditions are not met, and the session information cannot be accurately segmented. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes a variety of flexible session segmentation methods to meet different needs.

[0006] To achieve the above objectives, the present invention provides a session information segmentation method based on time intervals, comprising:

[0007] Obtain a sequence of behaviors of a target user, the sequence of behaviors including at least one behavior of the user triggered by a client and in chronological order;

[0008] The cutting rules include a preset time interval when cutting the behavior sequence, wherein the preset time interval is the maximum time interval between two adjacent behaviors;

[0009] The behavior sequence is segmented according to the target segmentation rule to obtain at least one session.

[0010] Furthermore, the step of segmenting the behavior sequence according to the target segmentation rule includes:

[0011] If a specified behavior is selected from the behavior sequence, and the time interval between the end of the previous behavior and the start of the next behavior is greater than or equal to the preset time interval, or if no other behavior occurs after the end of the previous behavior, then the previous behavior is taken as the end position of a new session. The specified behavior, the behavior at the end position, and the behavior between the specified behavior and the behavior at the end position are then divided into a session.

[0012] Furthermore, the specified behavior includes:

[0013] The first action of the action sequence; and / or

[0014] The first action following the last session that has been cut in the action sequence.

[0015] This invention also provides a method for segmenting session information based on hybrid rules, comprising:

[0016] Obtain a sequence of behaviors of a target user, the sequence of behaviors including at least one behavior of the user triggered by a client and in chronological order;

[0017] The cutting rules include a preset time interval, a start action, and an end action, wherein the preset time interval is the maximum time interval between two adjacent actions;

[0018] The behavior sequence is segmented according to the target segmentation rule to obtain at least one session.

[0019] Furthermore, the step of segmenting the behavior sequence according to the target segmentation rule includes:

[0020] The starting behavior or the first behavior in the behavior sequence is taken as the starting position of a session. If the time interval between the end of the previous behavior and the start of the next behavior is less than the preset time interval, and the ending behavior is found, then the found ending behavior, the behavior at the starting position, and the behavior between the found ending behavior and the behavior at the starting position are divided into a session.

[0021] Furthermore, the step of segmenting the behavior sequence according to the target segmentation rule further includes:

[0022] The first action after the end of the corresponding session is taken as the starting position of a new session. Starting from the starting position, if the time interval between the end of the previous action and the start of the next action is less than the preset time interval, and the ending action is found, then the found ending action, the action at the starting position, and the actions between the found ending action and the starting action are divided into a session.

[0023] Furthermore, if the time interval between the end of the previous action and the start of the next action is greater than or equal to the preset time interval, or if no other action occurs after the end of the previous action, then the previous action is taken as the end point of a new session. The step of segmenting the action sequence according to the target segmentation rule further includes:

[0024] The starting behavior, the ending behavior, and the behavior between the starting behavior and the ending behavior are divided into a session.

[0025] Furthermore, if the time interval between the end of the previous action and the start of the next action is less than the preset time interval, and the next starting action is found, then the actions before the next starting action are taken as the end position of the new session. Cutting the action sequence according to the target cutting rule further includes:

[0026] The starting behavior, the ending behavior, and the behavior between the starting behavior and the ending behavior are divided into a session.

[0027] The beneficial effects of this invention are:

[0028] This invention segments session information based on time intervals, making it easier to convert a user's continuous behavior into a session, balancing session integrity and time requirements, and better meeting the needs of session information analysis.

[0029] Meanwhile, this invention provides two methods for segmenting conversation information, which can match different conversation information segmentation rules to meet different scenario requirements, thereby completing different segmentation methods, achieving flexibility, and meeting the conversation information analysis needs of different products. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating the session information segmentation method based on time intervals according to an embodiment of the present invention.

[0031] Figure 2 This is a flowchart illustrating the session information segmentation method based on hybrid rules in an embodiment of the present invention.

[0032] Figure 3 This is a schematic diagram of session information segmentation provided in an embodiment of the present invention.

[0033] Figure 4 This is a schematic diagram of session information segmentation provided in another embodiment of the present invention.

[0034] Figure 5 This is a schematic diagram of session information segmentation provided in another embodiment of the present invention. Detailed Implementation

[0035] The present invention will be further explained below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0036] Before describing the solution of this invention, the relevant concepts involved in this invention will be explained.

[0037] Client: Software programs running on terminal devices, such as various apps (applications), browsers, etc. Terminal devices can include, but are not limited to, mobile phones, computers, etc.

[0038] A session is the overall sequence of user actions performed on a client-side application, arranged chronologically to achieve a specific goal (i.e., session segmentation). The segmentation method varies depending on the specific goal. For example, for a shopping client, if shopping is the specific goal, the user's actions between logging in and placing an order (including both actions) can be segmented into one session; if client usage is the specific goal, the user's actions between logging in and logging out (including both actions) can be segmented into one session.

[0039] The embodiments of the present invention will be described below, such as Figure 1 As shown, this invention provides a session information segmentation method based on time intervals, applied to electronic devices. The session segmentation method includes:

[0040] S101. Obtain the behavior sequence of the target user, the behavior sequence including at least one behavior triggered by the user based on the client and in chronological order;

[0041] In this embodiment of the application, the behavior sequence is a set of a series of behaviors performed by the target user on the client within a predetermined time period in chronological order, including at least one behavior.

[0042] Taking a shopping app as an example, after the client is launched, it includes actions such as entering the search page, searching for products, selecting products, entering recommendations, and exiting the search. Each of these actions is a single action event, and the sequence of actions can be all or part of the above actions.

[0043] S102. The cutting rule includes a preset time interval, which is the maximum time interval between two adjacent actions;

[0044] Specifically, the preset time interval is a threshold value representing the time interval between the end of one action and the start of the next, used to determine which actions are included in a session. That is, when segmenting a sequence of actions, if the time interval between the end of one action and the start of the next is greater than or equal to the preset time interval, then the previous action is considered the end point of a session. For example, if the preset time interval is 3 minutes, it means that in each segmented session, the time interval between the end of one action and the start of the next is less than 3 minutes, and the time interval between the end of the last action and the start of its next action is greater than or equal to 3 minutes.

[0045] S103. Cut the sequence according to the cutting rules to obtain at least one session.

[0046] Based on the above rules, the behavior sequence is segmented, including:

[0047] If a specified action is selected from the action sequence, and the time interval between the end of the previous action and the start of the next action is greater than or equal to a preset time interval, or if no other action occurs after the end of the previous action, then the previous action is taken as the end position of a new session. The specified action, the action at the end position, and the actions between the specified action and the action at the end position are then divided into a session.

[0048] Specifically, the specified action can be the first action in a sequence of actions; or it can be the first action after the last session that has been cut in the sequence of actions.

[0049] like Figure 3 As shown, assuming the preset time interval is 3 minutes, when segmenting a session, since the time interval between the end of behavior C and the start of behavior D is 3 minutes, which is equal to the preset time interval, behavior A to behavior C constitutes the first session. The next behavior, D, is the first behavior after the first session, and since no other behavior occurs after behavior E, behavior D to behavior E constitutes the second session.

[0050] like Figure 2 As shown, this embodiment of the invention also provides a session information segmentation method based on hybrid segmentation rules, including:

[0051] S201. Obtain the behavior sequence of the target user, the behavior sequence including at least one behavior triggered by the user based on the client and in chronological order;

[0052] Similar to the previous method embodiment, the behavior sequence is a set of a series of behaviors performed by the target user on the client within a predetermined time period, in chronological order, including at least one behavior.

[0053] Taking a shopping app as an example, after the client is launched, it includes actions such as entering the search page, searching for products, selecting products, entering recommendations, and exiting the search. Each of these actions is a single action event, and the sequence of actions can be all or part of the above actions.

[0054] S202, The cutting rules include a preset time interval, a start action, and an end action, wherein the preset time interval is the maximum time interval between two adjacent actions;

[0055] Specifically, the preset time interval is a threshold value representing the time interval between the end of the previous action and the start of the next action, used to determine which actions are included in the session. The start action can be used to define the start position of the session, and the end action can be used to define the end position of the session. Starting from the start action in the action sequence, if the time interval between the end of the previous action and the start of the next action is less than the preset time interval, and an end action is found, then the found end action, the start action, and the actions between the found end action and the start action are segmented into one session. If the time interval between the end of the previous action and the start of the next action is less than the preset time interval, and a next start action is found, then the actions before the next start action are taken as the end position of a new session, and the start action, the actions at the end position, and the actions between the start action and the end position are segmented into one session. If the time interval between the end of the previous action and the start of the next action is greater than or equal to the preset time interval, or if no other actions occur after the end of the previous action, then the previous action is taken as the end position of a new session, and the start action, the actions at the end position, and the actions between the start action and the end position are segmented into one session.

[0056] S203. Cut the sequence according to the cutting rules to obtain at least one session.

[0057] In some embodiments, segmenting the behavior sequence based on the above rules includes:

[0058] Starting from the initial behavior in the behavior sequence, if the time interval between the end of the previous behavior and the start of the next behavior is less than a preset time interval, and an ending behavior is found, then the found ending behavior, the initial behavior, and the behavior between the found ending behavior and the initial behavior are divided into a session.

[0059] like Figure 4 As shown, assuming a preset time interval of 3 minutes, behavior A is the starting behavior, and behavior B is the ending behavior. The first behavior in the behavior sequence is behavior A. Therefore, starting from the first behavior, since the time intervals of two adjacent behaviors are both less than the preset time interval, behavior A and behavior B are divided into the first session.

[0060] In some embodiments, after completing session segmentation, segmenting the behavior sequence based on the aforementioned target segmentation rules further includes:

[0061] The first action after the end of the corresponding session is taken as the starting position of a new session. Starting from the starting position, if the time interval between the end of the previous action and the start of the next action is less than the preset time interval, and an ending action is found, then the found ending action, the action at the starting position, and the actions between the found ending action and the starting action are divided into a session.

[0062] like Figure 4 As shown, after the first session is cut out, the first action (action C) after the first session is taken as the starting position of a new session. From this starting position to the next ending action (action B), since the time interval between two adjacent actions is less than the preset time interval, the second session is cut from action C to action B.

[0063] In some embodiments, if the time interval between the end of the previous action and the start of the next action is greater than or equal to a preset time interval, or if no other action occurs after the end of the previous action, the previous action is taken as the end point of a new session. Based on the above target segmentation rules, segmenting the action sequence further includes:

[0064] The start behavior, the end behavior, and the behavior between the start and end behaviors are divided into a session.

[0065] like Figure 5As shown, assuming a preset time interval of 3 minutes, behavior A is the starting behavior, and behavior B is the ending behavior. The first behavior in the behavior sequence is behavior A, so it starts from the first behavior. Since the time interval between the third and fourth behaviors is 3 minutes, the third behavior is the ending behavior, and the first behavior (behavior A) to the third behavior (behavior C) constitutes the first session. Continuing downwards, the sixth behavior (behavior A) is the start of the third session. After the seventh behavior ends, no other behaviors occur, so the seventh behavior is the end position of the new session. The sixth and seventh behaviors constitute the third session.

[0066] In some embodiments, if the time interval between the end of the previous action and the start of the next action is less than a preset time interval, and a next starting action is found, then the actions before the next starting action are taken as the end position of the new session. Based on the above target segmentation rules, segmenting the action sequence further includes:

[0067] The start behavior, the end behavior, and the behavior between the start and end behaviors are divided into a session.

[0068] like Figure 5 As shown, the first action (action D) after the first session ends serves as the starting point of a new session. This continues until the sixth action, where the time interval between adjacent actions is less than the preset time interval. However, because the sixth action is the starting action, the second session is divided from the fourth action (action D) to the fifth action (action B).

[0069] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for segmenting session information based on time intervals, characterized in that: Obtain a sequence of behaviors of a target user, the sequence of behaviors including at least one behavior of the user triggered by a client and in chronological order; The cutting rules include a preset time interval, a start action, and an end action, wherein the preset time interval is the maximum time interval between two adjacent actions; The behavior sequence is segmented according to the segmentation rules to obtain at least one session; including: The starting behavior or the first behavior in the behavior sequence is taken as the starting position of a session. If the time interval between the end of the previous behavior and the start of the next behavior is less than the preset time interval, and the ending behavior is found, then the found ending behavior, the behavior at the starting position, and the behavior between the found ending behavior and the behavior at the starting position are divided into a session. The first behavior after the end of the corresponding session is taken as the starting position of a new session. Starting from the starting position, if the time interval between the end of the previous behavior and the start of the next behavior is less than the preset time interval, and the ending behavior is found, then the found ending behavior, the behavior at the starting position, and the behavior between the found ending behavior and the behavior at the starting position are divided into a session. If the time interval between the end of the previous action and the start of the next action is greater than or equal to the preset time interval, or if no other action occurs after the end of the previous action, then the previous action is taken as the end point of a new session. The step of segmenting the action sequence according to the segmentation rule further includes: The starting behavior, the ending behavior, and the behavior between the starting behavior and the ending behavior are divided into a session; If the time interval between the end of the previous behavior and the start of the next behavior is less than the preset time interval, and the next starting behavior is found, then the behavior before the next starting behavior is taken as the end position of the new session. Cutting the behavior sequence according to the cutting rules also includes: cutting the starting behavior, the behavior at the end position, and the behavior between the starting behavior and the behavior at the end position into a session.

Images