Event data analysis method and device, storage medium and electronic equipment

Abstract

The invention discloses an event data analysis method and device, a storage medium and electronic equipment. The method comprises the steps that event sequence data generated by a target application program are obtained, a first event type is different from a second event type, a group of causal coefficients is determined according to the event sequence data, and the group of causal coefficients comprises a first event type and a second event type which are different from each other under the condition that an event corresponding to the first event type occurs; the target causal coefficient of the occurrence probability of the event corresponding to the second event type is determined according to the central position of the influence of the first event type on the second event type in time and the diffusion amplitude, and the target causal coefficient responds to the determined first event; and displaying a target event associated with the first event and the target influence coefficient according to the target causal coefficient. The technical problem that the analysis result is not accurate enough due to the fact that the analysis result of the event data is not fine in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of computers, and in particular, to an event data analysis method and device, a storage medium, and an electronic device. Background Art

[0002] At present, the analysis process of event data is generally determined by methods based on event correlation. However, the above-mentioned methods based on event correlation can only count the correlation between two events. There may be no influence relationship between the two events. In addition, the method only considers events with known influence, and the scope of influence that can be found is small. In actual scenarios, many types of events may have an impact, resulting in the analysis results of event data being not refined or accurate enough.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present application provide an event data analysis method and device, a storage medium, and an electronic device to at least solve the technical problem in the related art that the analysis results of event data are not precise, resulting in inaccurate analysis results.

[0005] According to one aspect of an embodiment of the present application, a method for analyzing event data is provided, comprising: acquiring event sequence data generated by a target application, wherein the event sequence data comprises first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data comprises the occurrence time of each event in the first group of events, and the second event sequence data comprises the occurrence time of each event in the second group of events; determining a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients comprises a target causal coefficient, and the target causal coefficient represents a condition where an event corresponding to the first event type has occurred. In this case, the probability of the event corresponding to the second event type occurring, the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position in time and the diffusion range of the influence of the first event type on the second event type, the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events; in response to determining the first event, the target event and the target influence coefficient associated with the first event are displayed according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0006] According to another aspect of an embodiment of the present application, a method for analyzing event data is also provided, comprising: obtaining event sequence data generated by a target application, wherein the event sequence data comprises first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data comprises an occurrence time of each event in the first group of events, and the second event sequence data comprises an occurrence time of each event in the second group of events; analyzing the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in an expected target account set using a target application, the conversion loss analysis interface comprises a predetermined key node associated with the conversion target event, the conversion loss analysis interface is used to represent the conversion loss data generated by an account in the target account set using the target application, the occurrence of the conversion target event in the conversion loss data represents a successful conversion, and the conversion target event The event did not occur, indicating that the conversion was not successful; in response to performing an interactive operation on the target interactive object, obtaining first event sequence data and second event sequence data from the event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes the event corresponding to the key node or the conversion target event, and the event type of the first event is the second event type; determining a target causal coefficient according to the first event sequence data and the second event sequence data, and displaying a target list according to the target causal coefficient, wherein the target causal coefficient represents the probability of the event corresponding to the second event type occurring when the event corresponding to the first event type has occurred, the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events, the target list includes a target event associated with the first event, a target influence coefficient, and an association relationship between the target event and the target influence coefficient, the event type of the target event is the first event type, the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event, and the target influence coefficient is determined by the target causal coefficient.

[0007] According to another aspect of an embodiment of the present application, there is also provided an event data analysis device, including: a first acquisition module, used to acquire event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes the occurrence time of each event in the first group of events, and the second event sequence data includes the occurrence time of each event in the second group of events; a first calculation module, used to determine a set of causal coefficients based on the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, and the target causal coefficient indicates that the event corresponding to the first event type has been the probability of occurrence of the event corresponding to the second event type in the event of occurrence, the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events; a first display module is used to display the target event and the target influence coefficient associated with the first event according to the target causal coefficient in response to the determination of the first event, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0008] Optionally, the device is used to determine a set of causal coefficients based on the event sequence data in the following manner: establish an initial excitation model; perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models; and determine the causal coefficients output by the multiple excitation models as the set of causal coefficients.

[0009] Optionally, the device is used to perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data in the following manner to obtain multiple excitation models: perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data in the following manner to obtain multiple excitation models, wherein the any two event types are regarded as the first event type and the second event type: input the first event sequence data and the second event sequence data into the initial excitation model round by round to obtain multiple likelihood estimation function output values ​​corresponding to multiple rounds of iterative training, wherein the initial excitation model inputs the first event sequence data and the second event sequence data in each round, and outputs a causal coefficient and a likelihood estimation function output value; determines a target likelihood estimation function output value with the maximum value based on the multiple likelihood estimation function output values; determines the model parameters of the initial excitation model that outputs the target likelihood estimation function output value as the model parameters of the target excitation model, and determines the target excitation model, wherein, when the likelihood estimation function output value output by the initial excitation model is not the maximum value, the model parameters of the initial excitation model are updated using the gradient descent method.

[0010] Optionally, the device is used to establish an initial excitation model in the following manner, including: using Gaussian functions with different centers to set a conditional intensity function for the initial excitation model, wherein the parameters of the conditional intensity function include a baseline parameter, an excitation parameter, and an attenuation parameter, the baseline parameter representing the probability of an event corresponding to the second event type occurring when the event corresponding to the first event type has not occurred, the excitation parameter representing the degree of causal influence of the first event type on the second event type, and the attenuation coefficient representing the decay rate of the influence of the first event type on the second event type over time; establishing the initial excitation model according to the conditional intensity function.

[0011] Optionally, the device is used to respond to determining a first event by displaying a target event and a target influence coefficient associated with the first event based on the target causal coefficient in the following manner: in response to determining the first event, integrating the target causal coefficient to determine a target integral value, wherein the integral calculation has a value range of 0 to positive infinity; and determining the target integral value as the target influence coefficient.

[0012] Optionally, the device is also used to: after determining a set of causal coefficients based on the event sequence data, obtain a pre-set filtering threshold, wherein the filtering threshold is used to screen out causal relationships between meaningless event types; delete causal coefficients in the set of causal coefficients whose values ​​are less than the filtering threshold; and determine causal coefficients in the set of causal coefficients whose values ​​are greater than or equal to the filtering threshold as the set of meaningful causal coefficients between event types.

[0013] Optionally, the device is further used to: analyze the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the expected target account set using a target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, the conversion loss analysis interface is used to represent conversion loss data generated by an account in the target account set using the target application, the occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion; in response to performing an interactive operation on a target interactive object, the first event sequence data and the second event sequence data are acquired, and a target list is displayed according to the first event sequence data and the second event sequence data, wherein the target interactive object is used to determine the first event, the first event at least includes an event corresponding to the key node or the conversion target event, and the target list includes the target event, the target influence coefficient, and the association between the target event and the target influence coefficient.

[0014] Optionally, the device is used to obtain the first event sequence data and the second event sequence data in response to performing an interactive operation on a target interactive object in the following manner, and display a target list based on the first event sequence data and the second event sequence data, including: in response to performing an interactive operation on a first interactive object, obtaining the first event sequence data and the second event sequence data, wherein the first interactive object is used to represent an edge connecting a first key node and a second key node, an event corresponding to the second key node is regarded as the first event, and the first key node is the parent node of the second key node; determining a first target event set based on the first event sequence data, the second event sequence data and the first key node, wherein the first target event set includes events occurring after the second event and before the first event, and the second event is an event corresponding to the first key node; determining each event in the first target event set as the target event, and displaying the target list.

[0015] Optionally, the device is used to obtain the first event sequence data and the second event sequence data in response to performing an interactive operation on a target interactive object, and display a target list based on the first event sequence data and the second event sequence data in the following manner: in response to performing an interactive operation on a second interactive object, obtain the first event sequence data and the second event sequence data, wherein the second interactive object is used to represent a second key node, and the event corresponding to the second key node is regarded as the first event; determine a second target event set based on the first event sequence data and the second event sequence data and the second key node, wherein the second target event set includes events occurring before the first event; determine each event in the second target event set as the target event, and display the target list.

[0016] Optionally, the device is used to obtain the first event sequence data and the second event sequence data in response to performing an interactive operation on a target interactive object, and display a target list based on the first event sequence data and the second event sequence data in the following manner: in response to performing an interactive operation on a third interactive object, obtain the first event sequence data and the second event sequence data, wherein the third interactive object is used to represent a third key node, and the event corresponding to the child node of the third key node is regarded as the first event; determine a third target event set based on the first event sequence data and the second event sequence data and the third key node, wherein the third target event set includes events occurring after the first event; determine each event in the third target event set as the target event, and display the target list.

[0017] According to another aspect of the embodiment of the present application, there is also provided an event data analysis device, comprising: a second acquisition module, for acquiring event sequence data generated by a target application, wherein the event sequence data comprises first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data comprises an occurrence time of each event in the first group of events, and the second event sequence data comprises an occurrence time of each event in the second group of events; a second calculation module, for analyzing the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in a desired target account set using a target application, the conversion loss analysis interface comprises a predetermined key node associated with the conversion target event, the conversion loss analysis interface is used to represent the conversion loss data generated by an account in the target account set using the target application, the conversion target event in the conversion loss data having occurred represents a successful conversion, and the conversion target event The event has not occurred, indicating that the conversion has not been successful; a third acquisition module, for, in response to performing an interactive operation on the target interactive object, acquiring first event sequence data and second event sequence data from the event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes the event corresponding to the key node or the conversion target event, and the event type of the first event is the second event type; a second display module, for determining a target causal coefficient according to the first event sequence data and the second event sequence data, and displaying a target list according to the target causal coefficient, wherein the target causal coefficient represents the probability of the event corresponding to the second event type occurring when the event corresponding to the first event type has occurred, the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events, the target list includes a target event associated with the first event, a target influence coefficient, and an association relationship between the target event and the target influence coefficient, the event type of the target event is the first event type, the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event, and the target influence coefficient is determined by the target causal coefficient.

[0018] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned event data analysis method when running.

[0019] According to another aspect of the embodiment of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above event data analysis method.

[0020] According to another aspect of the embodiments of the present application, there is also provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the event data analysis method through the computer program.

[0021] In an embodiment of the present application, event sequence data generated by a target application is obtained, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes the occurrence time of each event in the first group of events, and the second event sequence data includes the occurrence time of each event in the second group of events; a group of causal coefficients is determined based on the event sequence data, wherein a group of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when an event corresponding to the first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion amplitude of the influence of the first event type on the second event type in time, and the central position and diffusion amplitude are determined by the probability of occurrence of an event corresponding to the second event type in the first group of events. The occurrence time of the first event and the occurrence time of each event in the second group of events are determined; in response to determining the first event, the target event associated with the first event and the target influence coefficient are displayed according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event. By more accurately mining the causal relationship between events, multiple Gaussian functions are used to represent the incentive function, and the causal relationship between events is modeled more flexibly, so as to more accurately mine the causal relationship between events. By calculating the causal relationship strength matrix between all event pairs and setting a threshold to screen out meaningful causal relationships, related events can be automatically recommended to customers, thereby achieving the purpose of reducing the difficulty and workload of analyzing account behavior data, thereby achieving the technical effect of improving analysis efficiency, optimizing analysis results, and improving analysis accuracy, thereby solving the technical problem that the analysis results of event data in related technologies are not precise, resulting in inaccurate analysis results.

[0022] In addition, by discovering and recommending the reasons for user conversion or loss, targeted product optimization can be carried out to improve product conversion rate and user experience.

[0023] On the other hand, by preprocessing and screening the data, and the small amount of computation required for model training itself, efficient computing performance is achieved while ensuring analysis accuracy, and real-time analysis can be performed on large-scale account behavior data. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0025] Figure 1 is a schematic diagram of an application environment of an optional event data analysis method according to an embodiment of the present application;

[0026] Figure 2 is a flow chart of an optional event data analysis method according to an embodiment of the present application;

[0027] Figure 3 is a schematic diagram of an optional event data analysis method according to an embodiment of the present application;

[0028] Figure 4 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0029] Figure 5 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0030] Figure 6 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0031] Figure 7 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0032] Figure 8 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0033] Fig. 9 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0034] Fig.10 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0035] Fig.11 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0036] Fig.12 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application;

[0037] Fig.13 is a flowchart of another optional event data analysis method according to an embodiment of the present application;

[0038] Fig.14 is a schematic structural diagram of an optional event data analysis device according to an embodiment of the present application;

[0039] Fig.15 is a schematic structural diagram of an optional event data analysis device according to an embodiment of the present application;

[0040] Fig.16 is a schematic diagram of the structure of an optional event data analysis product according to an embodiment of the present application;

[0041] Fig.17 It is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0043] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0044] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following interpretation:

[0045] Experience Analysis: A SaaS product that provides services to mini-program developers. It collects events such as mini-program users’ operations and UI changes to provide customers with analysis capabilities related to the mini-program experience.

[0046] Conversion analysis: An analysis method in the experience analysis function that helps customers discover key nodes for improving products or reasons for churn.

[0047] Hawkes Process: A self-exciting random point process in which the appearance of a point increases the probability of the appearance of subsequent points.

[0048] Event sequence data: Sequence data consisting of a series of events with timestamps.

[0049] The present application is described below in conjunction with embodiments:

[0050] According to one aspect of an embodiment of the present application, a method for analyzing event data is provided. Optionally, in this embodiment, the method for analyzing event data can be applied to: Figure 1 In the hardware environment composed of the server 101 and the terminal device 103 shown in FIG. Figure 1 As shown, the server 101 is connected to the terminal 103 via a network, and can be used to provide services for the terminal device or the application installed on the terminal device. The application can be a video application, an instant messaging application, a browser application, an educational application, a game application, etc. A database 105 may be set up on the server or independently of the server to provide data storage services for the server 101, for example, a game data storage server. The above-mentioned network may include, but is not limited to, a wired network and a wireless network, wherein the wired network includes, a local area network, a metropolitan area network and a wide area network; the wireless network includes, Bluetooth, WIFI and other networks that implement wireless communication; the terminal device 103 may be a terminal configured with an application, and may include, but is not limited to, at least one of the following: a mobile phone (such as an Android phone, an iOS phone, etc.), a laptop computer, a tablet computer, a PDA, a MID (Mobile Internet Devices), a PAD, a desktop computer, a smart TV, an intelligent voice interaction device, a smart home appliance, a vehicle-mounted terminal, an aircraft, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a mixed reality (MR) terminal and other computer devices; the above-mentioned server may be a single server, a server cluster consisting of multiple servers, or a cloud server.

[0051] Combination Figure 1 As shown, the above-mentioned event data analysis method can be implemented in the terminal device 103 through the following steps:

[0052] S1, acquiring event sequence data generated by a target application on a terminal device 103, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0053] S2, determining a set of causal coefficients on the terminal device 103 according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, and the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events;

[0054] S3, in response to determining the first event on the terminal device 103, the target event associated with the first event and the target influence coefficient are displayed according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0055] Optionally, in this embodiment, the above event data analysis method can also be implemented by a server, for example, Figure 1 It is implemented in the server 101 shown; or it is implemented by the terminal device and the server together.

[0056] The above is only an example and is not specifically limited in this embodiment.

[0057] Optionally, as an optional implementation, as Figure 2 As shown, the analysis method of the above event data includes:

[0058] S202, acquiring event sequence data generated by the target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0059] Optionally, in this embodiment, the target application may be understood as an application that currently or in the future allows acquisition of various events generated during use, which may include but is not limited to application clients, applets, web pages, etc.

[0060] Optionally, in this embodiment, the event sequence data generated by the target application may include, but is not limited to, event sequence data corresponding to various events generated by triggering operations during the operation of the target application.

[0061] It should be noted that the events corresponding to the above event sequence data may include but are not limited to events corresponding to a click operation performed on a control on the mini program, a drag operation performed on a control on the application client, a page sliding operation performed on the web page interface, etc.

[0062] For example, Figure 3 is a schematic diagram of an optional event data analysis method according to an embodiment of the present application, such as Figure 3 As shown, in the live broadcast application, when the interactive operation performed by the viewing user on the subscription button is obtained, a corresponding event will be generated, and the corresponding event data can be obtained by analyzing the application.

[0063] It should be noted that when an application triggers the generation of multiple events, event data corresponding to the multiple events will be obtained, and the event data corresponding to the multiple events arranged according to timestamps is the above-mentioned event sequence data.

[0064] In an exemplary embodiment, the above-mentioned event sequence data can be understood as sequence data composed of a series of events with timestamps. This type of data is widely available. For example, in "Xiao Ming gets up at 8 o'clock, goes to school at 9 o'clock, and eats at 12 o'clock", getting up, going to school, and eating are three different types of events. Their occurrence in sequence constitutes an event sequence, which is expressed as [getting up->going to school->eating].

[0065] Optionally, in this embodiment, the first group of events and the second group of events can be understood as two groups of events under different event types, wherein the event types may include but are not limited to click events, scroll events, etc., that is, taking the first event type being a click event and the second group of events being a scroll event as an example, the first group of events includes multiple click events generated at different times, and the second group of events includes multiple scroll events generated at different times. The first event sequence data is the event sequence data determined by the event data corresponding to the multiple click events, and the second event sequence data is the event sequence data determined by the event data corresponding to the multiple scroll events.

[0066] It should be noted that the above-mentioned first event sequence data includes the occurrence time of each event in the first group of events, and the second event sequence data includes the occurrence time of each event in the second group of events. It can be understood that each event data is arranged in the event sequence data according to the timestamp corresponding to the occurrence time. Therefore, the occurrence timestamp of each event is obtained from the above-mentioned event sequence data.

[0067] In an exemplary embodiment, through preprocessing, all event sets can be obtained: E = e1, e2, ..., e n , where there are n different types of events, such as click, scroll, etc. The account behavior sequence can be represented as a list containing timestamps and event types: Among them, t j represents the time when the jth event occurs, Indicates that the type of the j-th event is i.

[0068] S204, determining a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, and the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events;

[0069] Optionally, in this embodiment, each causal coefficient in the above-mentioned set of causal coefficients is used to indicate the causal relationship between two different types of events in the above-mentioned event sequence data. Taking the above-mentioned two different types of events including a first event type and a second event type as an example, the target causal coefficient corresponding to the first event type and the second event type represents the probability of occurrence of the event corresponding to the second event type when the event corresponding to the first event type has occurred.

[0070] In an exemplary embodiment, the target causal coefficient may be understood as the probability that an event in the second group of events will occur after an event in the first group of events occurs.

[0071] For example, Figure 4 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Figure 4 As shown, in the live broadcast application, when the interactive operations performed by the viewing user on the subscription button and the scrolling operations performed on the live broadcast interface are obtained, corresponding events will be generated. The corresponding event data can be obtained by analyzing the application. At this time, taking the first group of events as click events and the second group of events as scrolling events as examples, the probability of the scrolling event at each time point after the next click event occurs can be calculated based on the occurrence timestamps of each click event in the current first group of events and the occurrence timestamps of each scrolling event in the second group of events.

[0072] It should be noted that the above target causal coefficient can be determined in the following way:

[0073] For each pair of event types (e i ,e j ), the activation function between two events is defined using the Hawkes point process as:

[0074]

[0075] Among them, λ ij (t) is the event type e j At time t, an event type e is received i The conditional intensity function of the influence can be used to determine the corresponding causal coefficient in the above set of causal coefficients, which is expressed as i In the event of j Probability of occurrence; μ ij is the baseline intensity, indicating the absence of e i In the event of j Probability of occurrence; α ij is the excitation coefficient, indicating e i For j The degree of causal influence of ij is the attenuation coefficient, which means e i For j The rate at which the effect decays over time.

[0076] It should be noted that after an event occurs, its impact on another event may take some time to be reflected. For example, the add-to-cart event and the purchase event. Therefore, a set of (L) Gaussian functions with different centers can be used to represent the above activation function to better capture the complex causal relationship between events. For each pair of event types (e i ,e j ), the conditional intensity function λ can be ij (t) is amended to read:

[0077]

[0078] Where L is the number of normal distributions; α ijl is the excitation coefficient of the lth normal distribution, indicating that e i For j The degree of causal influence; μ ijl is the mean of the lth normal distribution, indicating that e i For j The central position of the influence of ijl is the standard deviation of the lth normal distribution, indicating that e i For j By introducing multiple normal distributions, we can more flexibly model the causal relationship between events, for example, capturing interactions and non-monotonic effects at multiple time scales.

[0079] S206, in response to determining the first event, displaying the target event associated with the first event and the target influence coefficient according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0080] Optionally, in this embodiment, the first event may be determined by analyzing interactive operations obtained in the application. When it is necessary to analyze event data of the target application, the first event may be determined through interactive operations such as clicking and long pressing.

[0081] It should be noted that the above target event is determined based on the first event. The target event refers to an event that occurs before the first event. The above target impact coefficient refers to the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0082] For example, Figure 5 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Figure 5As shown, it includes a display interface for the analysis application, wherein a conversion loss tree displays multiple events associated with the target application and the order in which the events occur. By performing interactive operations on event identifiers in the conversion loss tree (for example, event 1, event 2, ..., event 5), the event corresponding to the event identifier selected by the interactive operation can be determined as the above-mentioned first event.

[0083] Of course, the edge between events may also be selected through interactive operations to determine the parent node connected by the edge as the first event.

[0084] Optionally, in this embodiment, the target impact coefficient indicates the degree of influence of the occurrence of the target event on the occurrence of the first event, that is, one or more target events that caused the occurrence of the first event and their corresponding degree of influence are displayed to the account using the analysis application.

[0085] For example, Figure 6 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Figure 6 As shown, the conversion loss tree displayed by the analysis application, when the edge between "page / list / list" and "open business details" is selected through interactive operation, the analysis content for the edge is displayed, including attribution analysis, wherein the attribution analysis records the related target events "click filter", "Toast displays 'coupon successful'", "onShow" with "open business details" as the first event, and their corresponding impact coefficients, which are the above-mentioned target impact coefficients.

[0086] It should be noted that the influence coefficient of 0.8 means that the influence of the "click to filter" event on the "open product details" event is 0.8. Similarly, the influence of each target event on the "open product details" event can be determined.

[0087] Through this embodiment, the event sequence data generated by the target application is obtained, wherein the event sequence data includes first event sequence data corresponding to the first group of events and second event sequence data corresponding to the second group of events, the event type of the first group of events is the first event type, the event type of the second group of events is the second event type, the first event type is different from the second event type, the first event sequence data includes the occurrence time of each event in the first group of events, and the second event sequence data includes the occurrence time of each event in the second group of events; a group of causal coefficients is determined according to the event sequence data, wherein a group of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of the event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion amplitude of the influence of the first event type on the second event type in time, and the central position and diffusion amplitude are determined by the probability of occurrence of the event corresponding to the first event type in the first group of events. The occurrence time and the occurrence time of each event in the second group of events are determined; in response to determining the first event, the target event associated with the first event and the target influence coefficient are displayed according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event. By more accurately mining the causal relationship between events, multiple Gaussian functions are used to represent the incentive function, and the causal relationship between events is modeled more flexibly, so as to more accurately mine the causal relationship between events. By calculating the causal relationship strength matrix between all event pairs and setting a threshold to screen out meaningful causal relationships, related events can be automatically recommended to customers, thereby reducing the difficulty and workload of analyzing account behavior data, thereby achieving the technical effect of improving analysis efficiency, optimizing analysis results, and improving analysis accuracy, thereby solving the technical problem that the analysis results of event data in related technologies are not precise, resulting in inaccurate analysis results.

[0088] As an optional solution, a set of causal coefficients is determined based on event sequence data, including: establishing an initial excitation model; performing multiple rounds of iterative training on the initial excitation model based on event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models; and determining the causal coefficients output by the multiple excitation models as a set of causal coefficients.

[0089] Optionally, in this embodiment, the initial incentive model may include but is not limited to for each pair of event types (e i ,e j ), the activation function between two events is defined using the Hawkes point process as:

[0090]

[0091] Among them, λ ij (t) is the event type e j At time t, an event type e is received i The conditional intensity function of the influence can be used to determine the corresponding causal coefficient in the above set of causal coefficients, which is expressed as i In the event of j Probability of occurrence; μ ij is the baseline intensity, indicating the absence of e i In the event of j Probability of occurrence; α ij is the excitation coefficient, indicating e i For j The degree of causal influence of ij is the attenuation coefficient, which means e i For j The rate at which the effect decays over time.

[0092] It should be noted that after an event occurs, its impact on another event may take some time to be reflected. For example, the add-to-cart event and the purchase event. Therefore, a set of (L) Gaussian functions with different centers can be used to represent the above activation function to better capture the complex causal relationship between events. For each pair of event types (e i ,e j ), the conditional intensity function λ can be ij (t) is amended to read:

[0093]

[0094] Where L is the number of normal distributions; α ijl is the excitation coefficient of the lth normal distribution, indicating that e i For j The degree of causal influence; μ ijl is the mean of the lth normal distribution, indicating that e i For j The central position of the influence of ijl is the standard deviation of the lth normal distribution, indicating that e i For j By introducing multiple normal distributions, we can more flexibly model the causal relationship between events, for example, capturing interactions and non-monotonic effects at multiple time scales.

[0095] For example, Figure 7is a schematic diagram of another optional event data analysis method according to an embodiment of the present application. The above-mentioned Hawkes point process can be as follows Figure 7 shown.

[0096] As an optional scheme, multiple rounds of iterative training are performed on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models, including: multiple rounds of iterative training are performed on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models, wherein any two event types are regarded as the first event type and the second event type: the first event sequence data and the second event sequence data are input into the initial excitation model round by round to obtain multiple likelihood estimation function output values ​​corresponding to the multiple rounds of iterative training, wherein the initial excitation model inputs the first event sequence data and the second event sequence data in each round, and outputs a causal coefficient and a likelihood estimation function output value; the target likelihood estimation function output value with the largest value is determined according to the multiple likelihood estimation function output values; the model parameters of the initial excitation model that outputs the target likelihood estimation function output value are determined as the model parameters of the target excitation model to determine the target excitation model, wherein, when the likelihood estimation function output value output by the initial excitation model is not the maximum value, the model parameters of the initial excitation model are updated using the gradient descent method.

[0097] In an exemplary embodiment, the maximum likelihood estimation (MLE) training model can be used. Given the account behavior sequence data, the goal is to maximize the likelihood function of the observed data. For the Hawkes point process model, the likelihood function can be expressed as:

[0098]

[0099] where Θ is the model parameter, including the baseline intensity μ ij , Excitation coefficient α ijl , mean μ ijl and standard deviation σ ijl To simplify the calculation, we can take the log-likelihood function and sum it:

[0100]

[0101] Among them, the first term can be directly obtained. Considering that the Gaussian function is used as the excitation function, the second term of the formula can also be directly obtained by the integral formula through the derivation method to increase the calculation efficiency.

[0102] It should be noted that since we need to maximize the log-likelihood function, we can use the gradient descent method to update the model parameters. After calculating these gradients, we use the gradient descent method to update the model parameters:

[0103]

[0104] Where η is the learning rate, is the gradient at the kth iteration. Through multiple iterations, the model parameters that maximize the log-likelihood function can be found.

[0105] As an optional solution, an initial excitation model is established, including: using Gaussian functions with different centers to set a conditional intensity function for the initial excitation model, wherein the parameters of the conditional intensity function include a baseline parameter, an excitation parameter, and an attenuation parameter, the baseline parameter represents the probability of an event corresponding to a second event type occurring when an event corresponding to the first event type has not occurred, the excitation parameter represents the degree of causal influence of the first event type on the second event type, the attenuation coefficient represents the decay rate of the influence of the first event type on the second event type over time, and the attenuation coefficient is determined according to the central position in time and the diffusion amplitude of the influence of the first event type on the second event type; the initial excitation model is established according to the conditional intensity function.

[0106] In an exemplary embodiment, after the Hawkes point process model is established, the causal relationship between two events can be measured by calculating the integral of the activation function from 0 to positive infinity. i ,e j ), the causal strength C can be calculated ij :

[0107]

[0108] λ ij Substituting (t) into the above formula, we get:

[0109]

[0110] Where T is the length of the observation time window (pre-set time window). Since the integral of the Gaussian function has an analytical solution, the above formula can be further simplified:

[0111]

[0112] where Φ(·) is the cumulative distribution function of the standard normal distribution.

[0113] As an optional scheme, in response to determining the first event, the target event and the target influence coefficient associated with the first event are displayed according to the target causal coefficient, including: in response to determining the first event, integrating the target causal coefficient to determine the target integral value, wherein the integral calculation value range is 0 to positive infinity; and determining the target integral value as the target influence coefficient.

[0114] As an optional scheme, after determining a set of causal coefficients based on the event sequence data, the above method also includes: obtaining a pre-set filtering threshold, wherein the filtering threshold is used to filter out causal relationships between meaningless event types; deleting causal coefficients in a set of causal coefficients whose values ​​are less than the filtering threshold; and determining causal coefficients in a set of causal coefficients whose values ​​are greater than or equal to the filtering threshold as a set of meaningful causal coefficients between event types.

[0115] In an exemplary embodiment, the causal relationship strength matrix C between all event pairs can be obtained by using the set of causal coefficients determined above. ij In order to extract meaningful causal relationships, a threshold τ (corresponding to the aforementioned filtering threshold) can be set to retain only those event pairs whose causal relationship strength is greater than τ:

[0116] C ′ ={(e i ,e j )∣C ij >τ}

[0117] Finally, the extracted causal relationships can be displayed in the corresponding panel in a list format.

[0118] As an optional solution, the above method also includes: analyzing the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the expected target account set using a target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, and the conversion loss analysis interface is used to represent the conversion loss data generated by the account in the target account set using the target application, and the occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion; in response to performing an interactive operation on a target interactive object, obtaining first event sequence data and second event sequence data, and displaying a target list according to the first event sequence data and the second event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to a key node or a conversion target event, and the target list includes a target event, a target influence coefficient, and an association between the target event and the target influence coefficient.

[0119] Optionally, in this embodiment, the above-mentioned conversion target events may include but are not limited to events corresponding to predetermined conversion targets. Taking the e-commerce platform as an example, the above-mentioned conversion target events can be understood as clicking a payment button, adding to a shopping cart, etc. Taking the live broadcast platform as an example, the above-mentioned conversion target events can be understood as clicking a follow button, sharing operations with others, etc.

[0120] Optionally, in this embodiment, the above-mentioned successful conversion can be understood as the above-mentioned conversion target event occurring when a certain account uses the above-mentioned target application, the above-mentioned conversion failure can be understood as the above-mentioned conversion target event not occurring when a certain account uses the above-mentioned target application, the above-mentioned key nodes represent event nodes that need to be displayed in the conversion loss analysis interface, in addition to the key nodes, there are also undisplayed target events between the event nodes, and the association relationship between the above-mentioned target influence coefficient and the target event can be understood as displaying the target influence coefficient and the target event in a one-to-one correspondence, for example, through page format arrangement, so that the corresponding target influence coefficient and the target event are displayed in the same row or column.

[0121] As an optional scheme, in response to performing an interactive operation on a target interactive object, first event sequence data and second event sequence data are obtained, and a target list is displayed based on the first event sequence data and the second event sequence data, including: in response to performing an interactive operation on a first interactive object, first event sequence data and second event sequence data are obtained, wherein the first interactive object is used to represent an edge connecting a first key node and a second key node, an event corresponding to the second key node is regarded as a first event, and the first key node is a parent node of the second key node; a first target event set is determined based on the first event sequence data, the second event sequence data and the first key node, wherein the first target event set includes events occurring after the second event and before the first event, and the second event is an event corresponding to the first key node; each event in the first target event set is determined as a target event, and a target list is displayed.

[0122] In an exemplary embodiment, Figure 8 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Figure 8 As shown, the first interactive object is the edge connecting event 1 and event 2 in the conversion loss tree. By performing an interactive operation on the interactive object, the first event sequence data corresponding to event 2 and the second event sequence data corresponding to event 1 can be obtained.

[0123] Among them, the event types corresponding to event 1 and event 2 can be determined first, and then the corresponding event sequence data can be obtained according to the relevant event types. The node corresponding to event 1 in the conversion loss tree is the above-mentioned first key node, and the node corresponding to event 2 in the conversion loss tree is the above-mentioned second key node.

[0124] It should be noted that the above event 1 is an example of the above second event, the above event 2 is an example of the above first event, and the events between the above event 1 and event 2 are the events in the above first target event set.

[0125] For example, Fig. 9is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Fig. 9 As shown, after each event in the first target event set is determined as a target event, a target list can be displayed, which includes each target event and the target influence coefficient corresponding to the target event, wherein event A, event B and event C are the aforementioned target events, that is, events occurring between event 1 and event 2.

[0126] As an optional scheme, in response to performing an interactive operation on a target interactive object, first event sequence data and second event sequence data are obtained, and a target list is displayed based on the first event sequence data and the second event sequence data, including: in response to performing an interactive operation on a second interactive object, first event sequence data and second event sequence data are obtained, wherein the second interactive object is used to represent a second key node, and the event corresponding to the second key node is regarded as the first event; a second target event set is determined based on the first event sequence data and the second event sequence data and the second key node, wherein the second target event set includes events that occur before the first event; each event in the second target event set is determined as a target event, and a target list is displayed.

[0127] In an exemplary embodiment, Fig.10 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Fig.10 As shown, the second interactive object is the node corresponding to event 1 in the conversion loss tree. By performing interactive operations on the interactive object, the first event sequence data corresponding to event 2 and the second event sequence data corresponding to event 1 and event 3 are obtained.

[0128] Among them, the event types corresponding to event 1, event 2 and event 3 can be determined first, and then the corresponding event sequence data can be obtained according to the relevant event types. The node corresponding to event 2 in the conversion loss tree is the second key node mentioned above.

[0129] It should be noted that the above event 2 is an example of the above first event, and the events between the above event 1 and event 2 and the events between the above event 2 and event 3 are the events in the above second target event set.

[0130] For example, Fig.11 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Fig.11As shown, after each event in the second target event set is determined as a target event, a target list can be displayed, which includes each target event and the target influence coefficient corresponding to the target event, wherein event A, event B and event C are the aforementioned target events, that is, events occurring between event 1 and event 2 and between event 2 and event 3.

[0131] As an optional scheme, in response to performing an interactive operation on a target interactive object, first event sequence data and second event sequence data are obtained, and a target list is displayed based on the first event sequence data and the second event sequence data, including: in response to performing an interactive operation on a third interactive object, first event sequence data and second event sequence data are obtained, wherein the third interactive object is used to represent a third key node, and events corresponding to child nodes of the third key node are regarded as first events; a third target event set is determined based on the first event sequence data and the second event sequence data and the third key node, wherein the third target event set includes events occurring after the first event; each event in the third target event set is determined as a target event, and a target list is displayed.

[0132] In an exemplary embodiment, Fig.12 is a schematic diagram of another optional event data analysis method according to an embodiment of the present application, such as Fig.12 As shown, the third interactive object is the node corresponding to event 1 in the conversion loss tree. By performing interactive operations on the interactive object, the first event sequence data corresponding to event 1 and the second event sequence data corresponding to event 2 are obtained.

[0133] Among them, the event types corresponding to event 1 and event 2 can be determined first, and then the corresponding event sequence data can be obtained according to the relevant event types. The node corresponding to event 1 in the conversion loss tree is the third key node mentioned above.

[0134] It should be noted that the above event 2 is an example of the above first event, and the events between the above event 1 and event 2 are the events in the above second target event set.

[0135] Optionally, as an optional implementation, as Fig.13 As shown, the analysis method of the above event data includes:

[0136] S1302, acquiring event sequence data generated by the target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0137] S1304, analyzing the event sequence data according to the predetermined conversion target event, generating a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the desired target account set using the target application, the conversion loss analysis interface includes predetermined key nodes associated with the conversion target event, and the conversion loss analysis interface is used to represent conversion loss data generated by an account in the target account set using the target application. If the conversion target event has occurred in the conversion loss data, it indicates that the conversion is successful, and if the conversion target event has not occurred, it indicates that the conversion is not successful;

[0138] S1306, in response to performing an interactive operation on a target interactive object, obtaining first event sequence data and second event sequence data from event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to a key node or a conversion target event, and the event type of the first event is a second event type;

[0139] S1308, determine the target causal coefficient based on the first event sequence data and the second event sequence data, and display the target list based on the target causal coefficient, wherein the target causal coefficient represents the probability of the event corresponding to the second event type occurring when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events. The target list includes the target event associated with the first event, the target influence coefficient, and the association between the target event and the target influence coefficient. The event type of the target event is the first event type. The target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event. The target influence coefficient is determined by the target causal coefficient.

[0140] The present application is further explained below with reference to specific examples:

[0141] This application uses the Hawkes point process to model the account behavior sequence, discovers the reasons for user conversion or loss by mining the model parameters, and recommends them to the users.

[0142] This application can be applied to, but is not limited to, the experience analysis module in the development platform. This product helps customers analyze the user experience on their mini-programs, discover anomalies (such as bugs, freezes, design problems), and improve conversions. Experience analysis provides various analysis services by collecting various events generated by users when using mini-programs. The data generated by each mini-program access constitutes an event sequence data.

[0143] Experience analysis provides a conversion analysis function module. Conversion analysis allows customers to set conversion goals (such as clicking a purchase button), and then the system automatically calculates the conversion path of users on the mini program from entry to the goal. Nodes represent events, edges represent conversions between events, and the width of edges and nodes represents the traffic of the path.

[0144] Customers can select nodes and edges on the conversion analysis graph to view the details panel corresponding to the node or edge. This application is applied to the details panel. Using this application, when a user clicks on a conversion edge, such as clicking on the edge between event A and event B, the system will recommend events that occurred after event A and before event B, which led to event B; when a user clicks on a node, the system will recommend events that may be caused by the event and events that led to the occurrence of the event. By recommending related events, this technology helps customers quickly discover potential optimization points for mini program products.

[0145] Technically, this application can be divided into the following steps:

[0146] S1, data screening and processing;

[0147] When the user selects a corresponding node or edge on the conversion graph, the system will pull the event sequence of conversion and loss involved in the node / edge from the database. In order to improve the calculation speed, only the event types with a frequency exceeding a certain threshold are retained.

[0148] Through preprocessing, we get the set of all events: E = e1, e2, …, e n There are n different types of events, such as click, scroll, etc. The account behavior sequence can be represented as a list containing timestamps and event types: Among them, t j represents the time when the jth event occurs, Indicates that the type of the j-th event is i.

[0149] S2, Hawkes point process modeling;

[0150] For each pair of event types (e i ,e j ), the activation function between two events is defined using the Hawkes point process as:

[0151]

[0152] Among them, λ ij (t) is the event type e j At time t, an event type e is received i The conditional intensity function of the influence can be used to determine the corresponding causal coefficient in the above set of causal coefficients, which is expressed as i In the event of j Probability of occurrence; μ ij is the baseline intensity, indicating the absence of e i In the event of j Probability of occurrence; α ij is the excitation coefficient, indicating e i For j The degree of causal influence of ij is the attenuation coefficient, which means e i For j The rate at which the effect decays over time.

[0153] The original activation function only uses an exponential function to simulate the impact between two events. In fact, when an event occurs, its impact on another event may take some time to be reflected. For example, the add-to-cart event and the purchase event. The activation function value described by the exponential function will only decrease as the number of events increases.

[0154] Therefore, a set of Gaussian functions with different centers are used to represent the activation function to better capture the complex causal relationship between events. i ,e j ), the conditional intensity function λ can be ij (t) is amended to read:

[0155]

[0156] Where L is the number of normal distributions; α ijl is the excitation coefficient of the lth normal distribution, indicating that e i For j The degree of causal influence; μ ijl is the mean of the lth normal distribution, indicating that e i For j The central position of the influence of ijl is the standard deviation of the lth normal distribution, indicating that e i For j By introducing multiple normal distributions, we can more flexibly model the causal relationship between events, for example, capturing interactions and non-monotonic effects at multiple time scales.

[0157] S3, model training;

[0158] Next, the model is trained using Maximum Likelihood Estimation (MLE). Given the account behavior sequence data, the goal is to maximize the likelihood function of the observed data. For the Hawkes point process model, the likelihood function can be expressed as:

[0159]

[0160] where Θ is the model parameter, including the baseline intensity μ ij , Excitation coefficient α ijl , mean μ ijl and standard deviation σ ijl To simplify the calculation, we can take the log-likelihood function and sum it:

[0161]

[0162] Among them, the first term can be directly obtained. Considering that the Gaussian function is used as the excitation function, the second term of the formula can also be directly obtained by the integral formula through the derivation method to increase the calculation efficiency.

[0163] We need to maximize this log-likelihood function. To do this, we can use gradient descent to update the model parameters.

[0164] After calculating these gradients, we can update the model parameters using gradient descent:

[0165]

[0166] Where η is the learning rate, is the gradient at the kth iteration. Through multiple iterations, the model parameters that maximize the log-likelihood function can be found.

[0167] S4, causal relationship extraction;

[0168] After establishing the Hawkes point process model, the causal relationship between two events can be measured by calculating the integral of the activation function from 0 to positive infinity. i ,e j ), the causal strength C can be calculated ij :

[0169]

[0170] λ ij Substituting (t) into the above formula, we get:

[0171]

[0172] Where T is the length of the observation time window (pre-set time window). Since the integral of the Gaussian function has an analytical solution, the above formula can be further simplified:

[0173]

[0174] where Φ(·) is the cumulative distribution function of the standard normal distribution.

[0175] Now, we have obtained the causal relationship strength matrix C between all event pairs = {C ij In order to extract meaningful causal relationships, a threshold τ (corresponding to the aforementioned filtering threshold) can be set to retain only those event pairs whose causal relationship strength is greater than τ:

[0176] C ′ ={(e i ,e j )∣C ij >τ}

[0177] S5, causal relationship display;

[0178] Finally, the extracted causal relationships are displayed in the corresponding panel in a list format.

[0179] Through the embodiments of the present application, the causal relationship between events can be mined more accurately: by using multiple Gaussian functions to represent the activation function, the causal relationship between events can be modeled more flexibly, such as capturing interactions and non-monotonic effects on multiple time scales. This enables more accurate mining of the causal relationship between events, thereby helping customers find potential optimization points.

[0180] On the other hand, the embodiment of the present application can automatically recommend related events: by calculating the causal relationship strength matrix between all event pairs and setting a threshold to screen out meaningful causal relationships, related events can be automatically recommended to customers. This greatly reduces the difficulty and workload of customers analyzing account behavior data and improves analysis efficiency.

[0181] On the other hand, the embodiments of the present application can improve the product optimization effect. By discovering and recommending the reasons for user conversion or loss, customers can optimize products in a targeted manner, thereby improving the product conversion rate and user experience. This will directly lead to higher customer satisfaction and commercial value.

[0182] On the other hand, the embodiment of the present application has efficient computing performance. By preprocessing and screening the data, and the small amount of computation in the model training itself, the present application achieves efficient computing performance while ensuring the accuracy of the analysis. This enables the present application to perform real-time analysis on large-scale account behavior data.

[0183] It is understandable that in the specific implementation of this application, related data such as user information is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0184] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0185] According to another aspect of the embodiment of the present application, there is also provided an event data analysis device for implementing the above-mentioned event data analysis method. Fig.14 As shown, the device comprises:

[0186] A first acquisition module 1402 is used to acquire event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0187] A first calculation module 1404 is used to determine a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to a second event type when an event corresponding to a first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, and the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events;

[0188] The first display module 1406 is used to display the target event associated with the first event and the target influence coefficient according to the target causal coefficient in response to determining the first event, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0189] As an optional solution, the device is used to determine a set of causal coefficients based on event sequence data in the following manner: establish an initial excitation model; perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models; and determine the causal coefficients output by the multiple excitation models as a set of causal coefficients.

[0190] As an optional solution, the device is used to perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data in the following manner to obtain multiple excitation models: perform multiple rounds of iterative training on the initial excitation model based on the event sequence data corresponding to any two event types in the event sequence data in the following manner to obtain multiple excitation models, wherein any two event types are regarded as the first event type and the second event type: input the first event sequence data and the second event sequence data into the initial excitation model round by round to obtain multiple likelihood estimation function output values ​​corresponding to the multiple rounds of iterative training, wherein the initial excitation model inputs the first event sequence data and the second event sequence data in each round, and outputs a causal coefficient and a likelihood estimation function output value; determine the target likelihood estimation function output value with the maximum value based on the multiple likelihood estimation function output values; determine the model parameters of the initial excitation model that outputs the target likelihood estimation function output value as the model parameters of the target excitation model, and determine the target excitation model, wherein, when the likelihood estimation function output value output by the initial excitation model is not the maximum value, the model parameters of the initial excitation model are updated using the gradient descent method.

[0191] As an optional scheme, the device is used to establish an initial excitation model in the following manner, including: using Gaussian functions with different centers to set a conditional intensity function for the initial excitation model, wherein the parameters of the conditional intensity function include a baseline parameter, an excitation parameter, and an attenuation parameter, the baseline parameter represents the probability of an event corresponding to a second event type occurring when an event corresponding to the first event type has not occurred, the excitation parameter represents the degree of causal influence of the first event type on the second event type, the attenuation coefficient represents the decay rate of the influence of the first event type on the second event type over time, and the attenuation coefficient is determined based on the central position in time and the diffusion amplitude of the influence of the first event type on the second event type; and establishing the initial excitation model based on the conditional intensity function.

[0192] As an optional scheme, the device is used to respond to the determination of the first event in the following manner, and display the target event associated with the first event and the target influence coefficient according to the target causal coefficient: in response to the determination of the first event, integrate the target causal coefficient to determine the target integral value, wherein the integral calculation value range is 0 to positive infinity; and determine the target integral value as the target influence coefficient.

[0193] As an optional scheme, the device is also used to: after determining a set of causal coefficients based on event sequence data, obtain a pre-set filtering threshold, wherein the filtering threshold is used to screen out meaningless causal relationships between event types; delete causal coefficients in a set of causal coefficients whose values ​​are less than the filtering threshold; and determine causal coefficients in a set of causal coefficients whose values ​​are greater than or equal to the filtering threshold as a set of meaningful causal coefficients between event types.

[0194] As an optional solution, the device is also used to: analyze the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the expected target account set using a target application, the conversion loss analysis interface includes a predetermined key node that has an association with the conversion target event, and the conversion loss analysis interface is used to represent the conversion loss data generated by the account in the target account set using the target application, and the occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion; in response to performing an interactive operation on a target interactive object, obtain first event sequence data and second event sequence data, and display a target list based on the first event sequence data and the second event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to a key node or a conversion target event, and the target list includes a target event, a target influence coefficient, and an association between the target event and the target influence coefficient.

[0195] As an optional scheme, the device is used to obtain first event sequence data and second event sequence data in response to performing an interactive operation on a target interactive object in the following manner, and display a target list based on the first event sequence data and the second event sequence data, including: in response to performing an interactive operation on a first interactive object, obtaining first event sequence data and second event sequence data, wherein the first interactive object is used to represent an edge connecting a first key node and a second key node, an event corresponding to the second key node is regarded as a first event, and the first key node is a parent node of the second key node; determining a first target event set based on the first event sequence data, the second event sequence data and the first key node, wherein the first target event set includes events occurring after the second event and before the first event, and the second event is an event corresponding to the first key node; determining each event in the first target event set as a target event, and displaying the target list.

[0196] As an optional scheme, the device is used to obtain first event sequence data and second event sequence data in response to performing an interactive operation on a target interactive object in the following manner, and display a target list based on the first event sequence data and the second event sequence data: in response to performing an interactive operation on a second interactive object, obtain first event sequence data and second event sequence data, wherein the second interactive object is used to represent a second key node, and the event corresponding to the second key node is regarded as the first event; determine a second target event set based on the first event sequence data and the second event sequence data and the second key node, wherein the second target event set includes events that occur before the first event; determine each event in the second target event set as a target event, and display the target list.

[0197] As an optional scheme, the device is used to obtain first event sequence data and second event sequence data in response to performing an interactive operation on a target interactive object in the following manner, and display a target list based on the first event sequence data and the second event sequence data: in response to performing an interactive operation on a third interactive object, obtain first event sequence data and second event sequence data, wherein the third interactive object is used to represent a third key node, and events corresponding to child nodes of the third key node are regarded as first events; determine a third target event set based on the first event sequence data and the second event sequence data and the third key node, wherein the third target event set includes events occurring after the first event; determine each event in the third target event set as a target event, and display the target list.

[0198] According to another aspect of the embodiment of the present application, there is also provided an event data analysis device for implementing the above-mentioned event data analysis method. Fig.15 As shown, the device comprises:

[0199] The second acquisition module 1502 is used to acquire event sequence data generated by the target application, wherein the event sequence data includes first event sequence data corresponding to the first group of events and second event sequence data corresponding to the second group of events, the event type of the first group of events is the first event type, the event type of the second group of events is the second event type, the first event type is different from the second event type, the first event sequence data includes the occurrence time of each event in the first group of events, and the second event sequence data includes the occurrence time of each event in the second group of events;

[0200] The second calculation module 1504 is used to analyze the event sequence data according to the predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the desired target account set using the target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, and the conversion loss analysis interface is used to represent the conversion loss data generated by the account in the target account set using the target application. In the conversion loss data, if the conversion target event has occurred, it means that the conversion is successful, and if the conversion target event has not occurred, it means that the conversion is not successful.

[0201] A third acquisition module 1506 is used to acquire first event sequence data and second event sequence data from the event sequence data in response to performing an interactive operation on a target interactive object, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to a key node or a conversion target event, and the event type of the first event is a second event type;

[0202] The second display module 1508 is used to determine the target causal coefficient based on the first event sequence data and the second event sequence data, and display the target list based on the target causal coefficient, wherein the target causal coefficient represents the probability of occurrence of the event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events. The target list includes the target event associated with the first event, the target influence coefficient, and the association relationship between the target event and the target influence coefficient. The event type of the target event is the first event type. The target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event, and the target influence coefficient is determined by the target causal coefficient.

[0203] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0204] According to one aspect of the present application, a computer program product is provided, the computer program product comprising a computer program / instruction, the computer program / instruction comprising a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 1609, and / or installed from a removable medium 1611. When the computer program is executed by the central processor 1601, various functions provided in the embodiments of the present application are executed.

[0205] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0206] Fig.16The structure block diagram of a computer system for implementing an electronic device according to an embodiment of the present application is schematically shown.

[0207] It should be noted that Fig.16 The computer system 1600 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0208] like Fig.16 As shown, the computer system 1600 includes a central processing unit 1601 (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory 1602 (ROM) or the program loaded from the storage part 1608 to the random access memory 1603 (RAM). Various programs and data required for system operation are also stored in the random access memory 1603. The central processing unit 1601, the read-only memory 1602 and the random access memory 1603 are connected to each other through a bus 1604. The input / output interface 1605 (Input / Output interface, i.e., I / O interface) is also connected to the bus 1604.

[0209] The following components are connected to the input / output interface 1605: an input section 1606 including a keyboard, a mouse, etc.; an output section 1607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1608 including a hard disk, etc.; and a communication section 1609 including a network interface card such as a LAN card, a modem, etc. The communication section 1609 performs communication processing via a network such as the Internet. A drive 1610 is also connected to the input / output interface 1605 as needed. A removable medium 1611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1610 as needed so that a computer program read therefrom is installed into the storage section 1608 as needed.

[0210] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1609, and / or installed from the removable medium 1611. When the computer program is executed by the central processor 1601, various functions defined in the system of the present application are executed.

[0211] According to another aspect of the embodiment of the present application, an electronic device for implementing the above-mentioned event data analysis method is also provided. The electronic device may be Figure 1 The terminal device or server shown in the figure. This embodiment is described by taking the electronic device as a terminal device as an example. Fig.17 As shown, the electronic device includes a memory 1702 and a processor 1704. The memory 1702 stores a computer program, and the processor 1704 is configured to execute the steps in any of the above method embodiments through the computer program.

[0212] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0213] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0214] S1, obtaining event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0215] S2, determining a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to a second event type when an event corresponding to a first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, and the central position and diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events;

[0216] S3, in response to determining the first event, displaying the target event associated with the first event and the target influence coefficient according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0217] Alternatively, a person skilled in the art may understand that: Fig.17 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, and other terminal devices. Fig.17 The electronic device and the electronic equipment described above are not limited in structure. Fig.17 More or fewer components (such as network interfaces, etc.) as shown in, or with Fig.17 Different configurations are shown.

[0218] Among them, the memory 1702 can be used to store software programs and modules, such as the program instructions / modules corresponding to the event data analysis method and device in the embodiment of the present application. The processor 1704 executes various functional applications and data processing by running the software programs and modules stored in the memory 1702, that is, realizing the above-mentioned event data analysis method. The memory 1702 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 1702 may further include a memory remotely located relative to the processor 1704, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. Among them, the memory 1702 can be specifically, but not limited to, used to store information such as event sequence data. As an example, such as Fig.17 As shown, the memory 1702 may include but is not limited to the first acquisition module 1402, the first calculation module 1404 and the first display module 1406 in the event data analysis device. In addition, it may also include but is not limited to other module units in the event data analysis device, which will not be repeated in this example.

[0219] Optionally, the transmission device 1706 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wired network and a wireless network. In one example, the transmission device 1706 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices and routers via a network cable so as to communicate with the Internet or a local area network. In one example, the transmission device 1706 is a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0220] In addition, the electronic device further includes: a display 1708 for displaying the target event and the target impact coefficient; and a connection bus 1710 for connecting the various module components in the electronic device.

[0221] In other embodiments, the terminal device or server may be a node in a distributed system, wherein the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting the multiple nodes through network communication. The nodes may form a peer-to-peer network, and any form of computing device, such as a server, terminal or other electronic device, may become a node in the blockchain system by joining the peer-to-peer network.

[0222] According to one aspect of the present application, a computer-readable storage medium is provided, and a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the event data analysis method provided in various optional implementations of the above-mentioned event data analysis aspects.

[0223] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0224] S1, obtaining event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events;

[0225] S2, determining a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to a second event type when an event corresponding to a first event type has occurred, and the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, and the central position and diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events;

[0226] S3, in response to determining the first event, displaying the target event associated with the first event and the target influence coefficient according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

[0227] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0228] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0229] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which can be personal computers, servers or network devices, etc.) to execute all or part of the steps of the methods of each embodiment of the present application.

[0230] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0231] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0232] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0233] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0234] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for analyzing event data, characterized in that: include: Acquire event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events; Determine a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of the event corresponding to the second event type occurring when the event corresponding to the first event type has occurred, the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion range of the influence of the first event type on the second event type in time, the central position and the diffusion range are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events; In response to determining a first event, a target event associated with the first event and a target influence coefficient are displayed according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

2. The method according to claim 1, characterized in that Determining a set of causal coefficients according to the event sequence data comprises: Establishing an initial incentive model; Based on the event sequence data corresponding to any two event types in the event sequence data, the initial excitation model is respectively trained for multiple rounds of iterations to obtain multiple excitation models; The causal coefficients output by the multiple excitation models are determined as the set of causal coefficients.

3. The method according to claim 2, characterized in that The event sequence data corresponding to any two event types in the event sequence data are used to perform multiple rounds of iterative training on the initial excitation model to obtain multiple excitation models, including: The initial excitation model is respectively subjected to multiple rounds of iterative training based on the event sequence data corresponding to any two event types in the event sequence data to obtain multiple excitation models, wherein the any two event types are regarded as the first event type and the second event type: Inputting the first event sequence data and the second event sequence data into the initial excitation model round by round to obtain a plurality of likelihood estimation function output values ​​corresponding to multiple rounds of iterative training, wherein the initial excitation model inputs the first event sequence data and the second event sequence data in each round and outputs a causal coefficient and a likelihood estimation function output value; Determine a target likelihood estimation function output value having a maximum value according to multiple likelihood estimation function output values; The model parameters of the initial excitation model that outputs the target likelihood estimation function output value are determined as the model parameters of the target excitation model, and the target excitation model is determined, wherein when the output value of the likelihood estimation function output by the initial excitation model is not a maximum value, the model parameters of the initial excitation model are updated using the gradient descent method.

4. The method according to claim 2, characterized in that: The establishing of the initial incentive model comprises: Using Gaussian functions with different centers to set a conditional intensity function for the initial excitation model, wherein the parameters of the conditional intensity function include a baseline parameter, an excitation parameter, and an attenuation parameter, the baseline parameter represents the probability of the event corresponding to the second event type occurring when the event corresponding to the first event type does not occur, the excitation parameter represents the degree of causal influence of the first event type on the second event type, and the attenuation coefficient represents the decay rate of the influence of the first event type on the second event type over time; The initial excitation model is established according to the conditional intensity function.

5. The method according to claim 1, characterized in that In response to determining the first event, displaying the target event associated with the first event and the target influence coefficient according to the target causal coefficient includes: In response to determining the first event, performing an integral calculation on the target causal coefficient to determine a target integral value, wherein the integral calculation value range is 0 to positive infinity; The target integral value is determined as the target influence coefficient.

6. The method according to claim 1, characterized in that After determining a set of causal coefficients according to the event sequence data, the method further comprises: Obtaining a preset filtering threshold, wherein the filtering threshold is used to filter out causal relationships between meaningless event types; Deleting the causal coefficients whose values ​​are less than the filtering threshold in the set of causal coefficients; The causal coefficients in the set of causal coefficients whose values ​​are greater than or equal to the filtering threshold are determined as the set of causal coefficients that are meaningful between event types.

7. The method according to claim 1, characterized in that The method further comprises: The event sequence data is analyzed according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated when an account in the desired target account set uses a target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, and the conversion loss analysis interface is used to represent conversion loss data generated when an account in the target account set uses the target application. The occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion. In response to performing an interactive operation on a target interactive object, the first event sequence data and the second event sequence data are obtained, and a target list is displayed based on the first event sequence data and the second event sequence data, wherein the target interactive object is used to determine the first event, the first event at least includes the event corresponding to the key node or the conversion target event, and the target list includes the target event, the target influence coefficient, and the association between the target event and the target influence coefficient.

8. The method according to claim 7, characterized in that In response to performing an interactive operation on a target interactive object, acquiring the first event sequence data and the second event sequence data, and displaying a target list according to the first event sequence data and the second event sequence data, comprises: In response to performing an interactive operation on a first interactive object, acquiring the first event sequence data and the second event sequence data, wherein the first interactive object is used to represent an edge connecting a first key node and a second key node, an event corresponding to the second key node is regarded as the first event, and the first key node is a parent node of the second key node; Determine a first target event set according to the first event sequence data, the second event sequence data and the first key node, wherein the first target event set includes events occurring after a second event and before the first event, and the second event is an event corresponding to the first key node; Each event in the first target event set is determined as the target event, and the target list is displayed.

9. The method according to claim 7, characterized in that: In response to performing an interactive operation on a target interactive object, acquiring the first event sequence data and the second event sequence data, and displaying a target list according to the first event sequence data and the second event sequence data, comprises: In response to performing an interactive operation on a second interactive object, acquiring the first event sequence data and the second event sequence data, wherein the second interactive object is used to represent a second key node, and an event corresponding to the second key node is regarded as the first event; Determine a second target event set according to the first event sequence data, the second event sequence data and the second key node, wherein the second target event set includes events occurring before the first event; Each event in the second target event set is determined as the target event, and the target list is displayed.

10. The method according to claim 7, characterized in that In response to performing an interactive operation on a target interactive object, acquiring the first event sequence data and the second event sequence data, and displaying a target list according to the first event sequence data and the second event sequence data, comprises: In response to performing an interactive operation on a third interactive object, acquiring the first event sequence data and the second event sequence data, wherein the third interactive object is used to represent a third key node, and an event corresponding to a child node of the third key node is regarded as the first event; Determine a third target event set according to the first event sequence data, the second event sequence data and the third key node, wherein the third target event set includes events occurring after the first event; Each event in the third target event set is determined as the target event, and the target list is displayed.

11. A method for analyzing event data, characterized in that: include: Acquire event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events; Analyze the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the desired target account set using a target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, and the conversion loss analysis interface is used to represent conversion loss data generated by an account in the target account set using the target application. The occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion. In response to performing an interactive operation on a target interactive object, obtaining first event sequence data and second event sequence data from the event sequence data, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to the key node or the conversion target event, and the event type of the first event is the second event type; A target causal coefficient is determined according to the first event sequence data and the second event sequence data, and a target list is displayed according to the target causal coefficient, wherein the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events. The target list includes target events associated with the first event, target influence coefficients, and an association between the target events and the target influence coefficients. The event type of the target event is the first event type. The target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event, and the target influence coefficient is determined by the target causal coefficient.

12. An event data analysis device, characterized in that: include: A first acquisition module is used to acquire event sequence data generated by a target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events; a first calculation module, configured to determine a set of causal coefficients according to the event sequence data, wherein the set of causal coefficients includes a target causal coefficient, the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when the event corresponding to the first event type has occurred, the target causal coefficient is determined by the decay rate of the influence of the first event type on the second event type over time, the decay rate is determined by the central position and diffusion amplitude of the influence of the first event type on the second event type in time, the central position and the diffusion amplitude are determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events; A first display module is used to, in response to determining a first event, display a target event associated with the first event and a target influence coefficient according to the target causal coefficient, wherein the event type of the first event is the second event type, the event type of the target event is the first event type, and the target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event.

13. An event data analysis device, characterized in that: include: A second acquisition module is used to acquire event sequence data generated by the target application, wherein the event sequence data includes first event sequence data corresponding to a first group of events and second event sequence data corresponding to a second group of events, the event type of the first group of events is a first event type, the event type of the second group of events is a second event type, the first event type is different from the second event type, the first event sequence data includes an occurrence time of each event in the first group of events, and the second event sequence data includes an occurrence time of each event in the second group of events; A second calculation module is used to analyze the event sequence data according to a predetermined conversion target event to generate a conversion loss analysis interface, wherein the conversion target event represents an event generated by an account in the desired target account set using a target application, the conversion loss analysis interface includes a predetermined key node associated with the conversion target event, the conversion loss analysis interface is used to represent conversion loss data generated by an account in the target account set using the target application, the occurrence of the conversion target event in the conversion loss data indicates a successful conversion, and the non-occurrence of the conversion target event indicates an unsuccessful conversion; A third acquisition module is configured to acquire first event sequence data and second event sequence data from the event sequence data in response to performing an interactive operation on a target interactive object, wherein the target interactive object is used to determine a first event, the first event at least includes an event corresponding to the key node or the conversion target event, and the event type of the first event is the second event type; A second display module is used to determine a target causal coefficient based on the first event sequence data and the second event sequence data, and display a target list based on the target causal coefficient, wherein the target causal coefficient represents the probability of occurrence of an event corresponding to the second event type when the event corresponding to the first event type has occurred, and the target causal coefficient is determined by the occurrence time of each event in the first group of events and the occurrence time of each event in the second group of events. The target list includes a target event associated with the first event, a target influence coefficient, and an association between the target event and the target influence coefficient. The event type of the target event is the first event type. The target influence coefficient represents the degree of influence of the occurrence of the target event on the occurrence of the first event, and the target influence coefficient is determined by the target causal coefficient.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program can be executed by a terminal device or a computer to execute the method described in any one of claims 1 to 11.

15. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method described in any one of claims 1 to 11 are implemented.

16. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 11 through the computer program.

Images