Method, device and program for use in data analysis

Abstract

To provide a method, a device and a computer readable storage medium used in data analysis.SOLUTION: A method used in data analysis includes: generating association between a history factor affecting a history event and a result of a history event generated by the history factor on the basis of at least history condition data associated with the history factor, and history result data indicating the result of the history event (510); determining at least one target result of interest, which is selected for a current event associated with the history event (520); determining a current factor affecting the current event, and at least one target object associated with at least the one target result of interest (530); and determining data of at least one target result of interest on the basis of at least the association, current condition data associated with the current factor, and at least the one target object (540).SELECTED DRAWING: Figure 5

Description

[Technical field]

[0001] TECHNICAL FIELD Embodiments of the present disclosure relate to the field of artificial intelligence, and more particularly, to methods, devices and storage media used in data analysis. [Background technology]

[0002] Generally, users want to evaluate the effectiveness of a decision or compare the effectiveness of various decision plans before making a decision. For example, in the field of marketing research, decision makers want to understand whether adjusting a product, service, brand, or other strategy can improve customer satisfaction, and which strategy will provide the best improvement.

[0003] For example, in the area of ​​human resource management, managers want to understand the incentive effects of various employee incentive policies in order to balance the interests of the company with those of their employees, and in the area of ​​fault detection, operators want to understand which technical processes, operating methods, and equipment configurations can be improved to reduce the likelihood of a fault.

[0004] The potential effects of such decisions can be predicted by a strategy prediction system based on a causal model. Such a causal model can usually be established by historical data. After the causal model is established, by inputting different strategies, the strategy prediction system can predict the effects of the strategies, evaluate the effects of decisions, and help the user select the strategy with the greatest effect. Summary of the Invention [Problem to be solved by the invention]

[0005] In embodiments of the present disclosure, a method, a device, and a storage medium for use in data analysis are provided. [Means for solving the problem]

[0006] In a first aspect of the present disclosure, a method for use in data analysis is provided, the method including: generating an association between historical factors and outcomes of the historical events based at least on historical condition data associated with historical factors that influenced the historical events and historical outcome data indicating outcomes of the historical events caused by the historical factors, determining at least one goal outcome of interest selected for a current event associated with the historical events, determining current factors that influence the current event and at least one goal object associated with the at least one goal outcome of interest, and determining data of the at least one goal outcome of interest based at least on the association, current condition data associated with the current factors, and the at least one goal object.

[0007] In a second aspect of the present disclosure, a device for use in data analysis is provided, the device includes at least one processing unit, and at least one memory coupled to the at least one processing unit, storing instructions executed by the at least one processing unit, which when executed by the at least one processing unit, cause the device to perform operations, the operations include: generating an association between historical factors and outcomes of the historical events based at least on historical condition data associated with historical factors that influenced the historical events and historical outcome data indicating outcomes of the historical events caused by the historical factors, determining at least one goal outcome of interest to be selected for a current event associated with the historical events, determining current factors that influence the current event and at least one goal object associated with the at least one goal outcome of interest, and determining data of the at least one goal outcome of interest based at least on the association, current condition data associated with the current factors, and the at least one goal object.

[0008] In a third aspect of the present disclosure, there is provided a computer readable storage medium having stored thereon computer readable program instructions for performing the method described in the first aspect.

[0009] The Summary of the Invention is intended to present a selection of concepts in a simplified manner, which will be further described in the following embodiments. The description in the Summary of the Invention is not intended to identify key or required features of the disclosure, nor is it intended to limit the scope of the disclosure. [Brief description of the drawings]

[0010] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the drawings, in which like reference numerals generally refer to like parts.

[0011] [Figure 1] FIG. 1 is a schematic diagram of an environment in which embodiments of the present disclosure can be implemented.

[0012] [Diagram 2] 1 is a directed acyclic graph of an example of causal relationships according to an embodiment of the present disclosure.

[0013] [Diagram 3] FIG. 1 is a schematic diagram of one use scenario according to an embodiment of the present disclosure.

[0014] [Figure 4] FIG. 1 is a schematic diagram of an interactive process of data analysis according to an embodiment of the present disclosure.

[0015] [Diagram 5] 1 is a flow chart of an exemplary process of data analysis according to an embodiment of the present disclosure.

[0016] [Figure 6] 1 is a flow chart of a process for determining an association between historical factors and historical event outcomes according to an embodiment of the present disclosure.

[0017] [Figure 7] 1 is a flow chart of a process for determining a target object according to an embodiment of the present disclosure.

[0018] [Figure 8] FIG. 1 is a block diagram of an exemplary device for implementing the teachings of the present disclosure; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Hereinafter, preferred embodiments of the present disclosure will be described in more detail with reference to the drawings. Although the drawings show preferred embodiments of the present disclosure, it should be understood that the present disclosure can be realized in various forms and should not be limited to the embodiments described herein. Rather, these embodiments are provided to make the present disclosure more complete and to comprehensively convey the scope of the present disclosure to those skilled in the art.

[0020] As used herein, the term "comprises" and variations thereof refer to the open "including, but not limited to." Unless otherwise specified, the term "or" refers to "and / or" and the term "based on" refers to "based at least in part on." The terms "one exemplary embodiment" and "an embodiment" refer to "at least one exemplary embodiment." The term "another embodiment" refers to "at least one other embodiment." The terms "first," "second," etc. can refer to different or the same object. Other explicit and implicit definitions may be included in the following text.

[0021] In the embodiments of the present disclosure, the term "model" generally refers to the relationship and structure of a system that is expressed in a general or approximate manner using mathematical language with respect to features that refer to the system. Generally, a model can be generated by training using known data. The generated model can include model configurations and model parameters, etc. The model parameters may vary depending on the specific type of model. The term "causal model" generally describes the causal effect structure of a system.

[0022] Generally, users want to evaluate the effect of a decision or compare the effect of various decision plans before making a decision. For example, in retail, decision makers want to understand how much revenue each promotion plan will bring before creating a promotion plan, so as to select the most suitable promotion plan.

[0023] With the continuous development of computer technology, data analysis has been widely applied in various aspects of people's lives. For example, analytical devices such as deep neural network (DNN) models are increasingly applied in various tasks such as decision-making evaluation and goal prediction. Therefore, the possible effects of the above-mentioned decisions can be predicted by establishing a causal model. Such a causal model can usually be established by historical data. After the establishment of the causal model, by inputting different strategies, the strategy prediction system can predict the effects of the strategies, thereby evaluating the effects of decision-making, and meeting the needs of users in various fields.

[0024] However, for data for generating causal models, traditional solutions require a specific type of data sample type, so there are certain limitations to model generation. In addition, causal models generated by traditional solutions, such as models generated based on Bayesian networks, have some problems in construction. For example, such causal models can only realize prediction of average effects, and cannot make more accurate predictions for individuals or different groups, so they cannot meet users' needs for refined decision-making. Therefore, it is desirable to make more accurate and tailored predictions for decision-making for target objects at different granularities.

[0025] It is also desirable to realize an automated data analysis scheme that can continuously or periodically obtain data samples as decision-making data from, for example, user terminals or information-gathering devices such as industrial sensors, and analyze the effectiveness of corresponding decisions according to customer requirements.

[0026] Furthermore, it may be desirable to evaluate the predicted effect of the decision and, if a discrepancy is found to exist between the predicted effect of the decision and the effect desired by the user, generate a warning to the user so that the user can adjust their strategy in a timely manner.

[0027] According to some embodiments of the present disclosure, a solution for data analysis is provided. In the solution, an association between a historical factor and a result of a historical event can be generated based at least on historical condition data associated with a historical factor that influenced the historical event and historical result data. The historical result data indicates the result of the historical event caused by the historical factor. Then, at least one goal result of interest can be determined to be selected for a current event. The current event has a certain relationship with the historical event. Then, a current factor that influences the current event and at least one goal object associated with the at least one goal result of interest can be determined, and data of the at least one goal result of interest can be determined based at least on the association, the current condition data associated with the current factor, and the at least one goal object. In this manner, more accurate prediction can be achieved, and a user's desire for refined decision-making can be met.

[0028] Example Environment

[0029] Hereinafter, the embodiments of the present disclosure will be described in detail with reference to the drawings. Fig. 1 shows a schematic diagram of an exemplary environment 100 in which multiple embodiments of the present disclosure can be implemented. As shown in Fig. 1, the exemplary environment 100 includes a data analysis device 110. The data analysis device 110 may include a database 111 and a prediction unit 112.

[0030] The database 111 can be used to store necessary data files acquired through multiple routes, such as information collection devices, questionnaire surveys, etc. For example, for a retail business, the database 111 can store information such as sales promotion plans and historical sales records.

[0031] The database 111 may also perform general pre-processing on the data, such as steps such as abnormal data detection, data cleansing, missing value imputation, sample filtering, factor selection, etc., to improve data quality.

[0032] The prediction unit 112 can be used to learn specific knowledge using existing data extracted from the database 111 to process new data. The prediction unit 112 can be designed to perform various tasks, such as decision evaluation, outcome prediction, target detection, etc. Examples of the analysis unit 120 include, but are not limited to, various deep neural networks (DNNs), convolutional neural networks (CNNs), support vector machines (SVMs), decision trees, random forest models, etc.

[0033] In some embodiments, when the prediction unit 112 is used to perform a decision-making evaluation task, such as evaluating the impact of a promotional event on sales revenue of a shopping mall based on some commercial data, the prediction unit 112 can provide an evaluation for factors associated with the input shopping mall promotional event.

[0034] In some embodiments, the prediction unit 112 can retrieve commercial data associated with historical promotional events from the database 111 and generate a causal model between factors that influenced the historical promotional events and outcomes (e.g., sales) based on this data. In some embodiments, factors that influence a current promotional event can be input into the causal model established by the analysis unit 120 to predict outcomes of interest associated with the current promotional event.

[0035] In some embodiments, when the prediction unit 112 is used to perform target detection in industrial automation scenarios, the prediction unit 112 analyzes, for example, data associated with technology flows, operation methods, and equipment configurations to analyze their impact on equipment failures.

[0036] In some embodiments, the prediction unit 112 can periodically obtain data samples associated with equipment operations and equipment attributes from an external information collection device, and perform data analysis of the effects of the data samples on equipment failure. If the prediction unit 112 determines that the likelihood of equipment failure caused by some equipment operations and equipment attributes exceeds a threshold range, it can further issue a warning to a user.

[0037] It should be appreciated that the operating environment 100 depicted in Figure 1 is for illustrative purposes only and is not intended to be a limiting example of the operating environment. The operating environment 100 may further include any number of computing or processing units.

[0038] The process of generating a causal model in the prediction unit 112 and making a prediction for a given goal will be described in more detail below in connection with FIG.

[0039] Examples of causal relationships

[0040] 2 illustrates a directed acyclic graph (DAG) of an example of causal relationships according to an embodiment of the present disclosure. Before describing the causal model generation and prediction process, one example of causal effects will first be described in detail with reference to FIG.

[0041] In general, a DAG is used to explain causal effects between multiple variables. The DAG may include nodes that represent variables and directed edges and paths that represent causal effects between the variables. For example, a directed edge from a parent node to its child node may represent a direct causal effect between the variable represented by the parent node and the variable represented by the child node. Also, for example, a path from one node to another node may represent an indirect causal effect between the variables represented by the two nodes.

[0042] 2, the variable "preferential treatment margin" 210 and the variable "product type" 230 can be considered as treatment variables, and the variable "sales promotion result" 240 can be considered as an outcome variable. Edges 203 and 204 from the variables "preferential treatment margin" 210 and "product type" 230 to the variable "sales promotion result" 240 can be defined as direct edges. This indicates that the variables "preferential treatment margin" 210 and "product type" 230 can directly affect the variable "sales promotion result" 240.

[0043] 2 further includes an intermediate variable "customer experience" 220. An edge 201 from the variable "preferential treatment margin" 210 to the variable "customer experience" 220 can be defined as a direct edge, while an edge 202 from the variable "preferential treatment margin" 210 to the variable "sales promotion result" 240 via the variable "customer experience" 220 can be defined as an indirect edge. This indicates that while the variable "preferential treatment margin" 210 can affect the variable "customer experience" 220, the variable "customer experience" 220 can also affect the variable "sales promotion result" 240.

[0044] For example, if the discount rate is relatively large, too many customers may come to the store to purchase products, which may exceed the load of the shopping mall. As a result, the shopping mall may have a shortage of parking lots, the service level of sales staff may decline, and inventory may be insufficient, which may affect the customer experience, and the customer experience may also affect the final sales.

[0045] It should be understood that the directed acyclic graph shown in Figure 2 is merely an example of the condition variables (i.e., process variables and parameters) that may affect the outcome variables. The directed acyclic graph shown in Figure 2 may also include other possible process variables and parameters, as well as direct and / or indirect edges between them and the outcome variables.

[0046] Example use cases

[0047] 3 is a schematic diagram of one application scenario according to an embodiment of the present disclosure. Hereinafter, the application according to the embodiment of the present disclosure will be further described with reference to FIG.

[0048] As shown in Fig. 3, a usage scenario 500 may include the data analysis device 110 and the user terminal 310 in Fig. 1. Communication is possible between the data analysis device 110 and the user terminal 310.

[0049] For example, in the causal model establishment phase, the data analysis device 110 can periodically obtain multiple data samples of condition data associated with a target outcome that may affect the target outcome to be predicted from the user terminal, thereby enabling the data analysis device 110 to establish a causal model for predicting the target outcome based on a large number of data samples.

[0050] In some embodiments, when the data analysis device 110 performs target detection in an industrial automation scene, the user terminal 310 can be connected to, for example, multiple industrial sensors (not shown) located in the industrial automation scene. The industrial sensors report collected environmental data to the user terminal 310. These environmental data are transmitted to the data analysis device 110 via the user terminal 310, so that the data analysis device can obtain sufficient data samples.

[0051] In addition, in the result prediction stage, the data analysis device 110 can obtain some individual requests from the user for the prediction process from the user terminal 310, such as prediction goals of interest, constraints on data samples, etc. After completing the prediction, the data analysis device 110 can transmit the prediction result to the user terminal 310.

[0052] In some embodiments, if the data analysis device 110 determines that a relatively large discrepancy exists between the predicted outcome and the outcome desired by the user, the data analysis device 110 may further send a warning to the user terminal 310 regarding the discrepancy.

[0053] It should be understood that Fig. 3 only shows an exemplary application scenario, which may include any number of user terminals 310. Also, the data analysis device 110 may be, for example, a module of the user terminal 310, or may be integrated in the user terminal 310 in chip form.

[0054] Example Interactive Process

[0055] Fig. 4 shows a schematic diagram of an interactive process 400 of data analysis according to an embodiment of the present disclosure. The interactive process can be implemented in the use scenario shown in Fig. 3. Therefore, the interactive process will be described again by taking the data analysis device 110 and the user terminal 310 in Fig. 3 as examples.

[0056] The data analysis device 110 can obtain condition data of factors affecting an event and data of the event result caused by the factors from the user terminal 310 (405). In some embodiments, the condition data can be periodically or continuously transmitted from the user terminal 310 to the data analysis device 110. The condition data can be, for example, data input by a user to the user terminal 310. The condition data can also be data collected by other data-collecting peripheral devices, such as sensors, connected to the user terminal 310 and reported to the user terminal 310.

[0057] The data analysis device 110 can determine 410 a causal relationship between the condition data and the result data using a large number of data samples. For example, a sensor collects multiple data samples of temperature data for the part, environmental data for the industrial environment in which the part is located, frequency of use data for the part, and aging data for the part. For example, the data analysis device 110 can determine, for the part, the impact of the part's temperature level, environmental factors, and frequency of use on the aging of the part. The data analysis device 110 can also further predict, for example, the likelihood of the part failing, the frequency interval at which the failure occurs, and the associated maintenance costs for the equipment.

[0058] The user terminal 310 may transmit (415) an instruction about the other event to the data analysis device 110. The instruction may include, for example, one or more of the following information: a predicted goal of interest to the other event by the user, possible factors affecting the other event, constraints on the possible factors, and a goal object associated with the goal outcome of interest to the user.

[0059] For example, the data analysis device 110 obtains, from the instruction, condition data of possible factors that affect the other event and a prediction goal that the user is interested in for the other event. For example, the condition data of the possible factors are some environmental parameters of the operating environment in which the other part is located, and the other part may be made of the same material as the part in the previous event. The data analysis device 110 can determine that the other event is associated with the previous event, and can then obtain a causal relationship model established based on the previous event.

[0060] The data analysis device 110 can determine (420) a predicted target outcome of interest to the user for the other event based on the causal model and one or more pieces of information obtained from an instruction by the user terminal 310. For example, the data analysis device 110 can input one or more pieces of information obtained from an instruction by the user terminal 310 into a causal model and set an output result of the causal model as a predicted target outcome.

[0061] The data analysis device 110 can transmit the result of the prediction target to the user terminal 310 (425). If the user is not satisfied with the result of the prediction target, the user can adjust the condition data of the factors affecting the event and transmit it to the data analysis device 110 to perform prediction again.

[0062] Also, the data analysis device 110 can perform an automatic analysis based on the predicted target result. For example, the user terminal 310 has previously transmitted a range of desired results to the data analysis device 110. If the data analysis device 110 determines based on the range of desired results that the difference between the predicted result and the desired result exceeds a threshold difference, it can send a warning about the difference to the user terminal.

[0063] The data analysis process according to the embodiment of the present disclosure has been briefly described above from the perspective of data interaction. The details of the data analysis will be further elaborated below in conjunction with another example.

[0064] Example process of data analysis

[0065] The process used for data analysis will now be described in more detail with reference to Figures 5-7. Figure 5 shows a flow chart of a process 500 used for data analysis according to some embodiments of the present disclosure. The process 500 may be implemented by the data analysis device 110 of Figure 1. For ease of discussion, the process 500 will be described with reference to Figure 1.

[0066] 5, in block 510, the data analysis device 110 generates an association between the historical factors and the outcomes of the historical events based at least on the historical condition data associated with the historical factors affecting the historical events and the historical outcome data, the historical outcome data indicating the outcomes of the historical events caused by the historical factors.

[0067] In some embodiments, the data analysis device 110 can obtain, for example, historical condition data. The historical condition data can be obtained, for example, from the database 111 shown in FIG. 1 . The historical condition data can also be obtained in other ways. For example, the historical condition data can be periodically obtained from a data collection device and transmitted to the data analysis device 110. As described above, the data collection device is directly or indirectly connected to the data analysis device 110. The data collection device can be, for example, a plurality of sensors.

[0068] In some embodiments, the historical condition data obtained may include different data types. For example, the data samples of the historical condition data may be discrete. Also, the data samples of the historical condition data may be continuous. Or, the data samples of the historical condition data may include continuous data samples as well as discrete data samples.

[0069] In some embodiments, the data analysis device 110 can further determine historical result data associated with the historical condition data. The reason is that the historical condition data is associated with historical factors that affect the historical event, and the historical result data can indicate the result of the historical event caused by the historical factors. For example, according to the contents shown in the directed acyclic graph representing the causal relationship shown in FIG. 2, the historical condition data may be, for example, the status of the preferential treatment range, and the historical result data may be, for example, the status of the sales promotion result. The historical result data may be, for example, directly obtained from the database 111. The historical result data may be, for example, obtained from historical event record information input from another external device.

[0070] The data analysis device 110 can generate associations between historical causes and outcomes of historical events based on the historical condition data and the historical outcome data.

[0071] One possible situation is that the historical result data associated with a certain historical condition data cannot be directly obtained, for example, because the number of samples of the data samples of the acquired historical condition data is insufficient or simple. For example, the data samples of the acquired historical condition data are data on cosmetics sales promotion, while the historical result is revenue on women's fashion sales promotion. Different product types have different consumption groups, consumption cycles, and consumption abilities, so a direct causal relationship between the historical condition data and the historical result data cannot be established.

[0072] The relationship between them may have to be determined using other data, and therefore block 510 of Figure 5 is further detailed below in conjunction with Figure 6. Figure 6 illustrates a flow chart of a process for determining associations between historical factors and historical event outcomes according to an embodiment of the present disclosure.

[0073] In block 610, the data analysis device 110 acquires historical condition data. In block 620, the data analysis device 110 determines whether the number of data samples of the acquired historical condition data is less than a threshold number. If the number of data samples of the historical condition data is less than the threshold number, i.e., the number of samples is insufficient, in block 630, the data analysis device 110 acquires reference data. The reference data includes, for example, expert knowledge. The reference data can indicate the degree of influence of the historical condition data on the historical result.

[0074] 6, in block 640, the data analysis device 110 may determine historical result data according to the historical condition data and the reference data. In block 650, the data analysis device 110 may generate the association based on the historical condition data and the historical result data.

[0075] Returning to block 620, if the number of data samples of the historical condition data is greater than the threshold number, the data analysis device 110 may determine historical result data based on the historical condition data in block 660, and generate the association based on the historical condition data and the historical result data as described above in block 650. It should be understood that even if the samples of the historical condition data are sufficient, expert knowledge may be obtained to further optimize the association establishment process.

[0076] In addition, after generating the association between the factors that affect the historical events and the historical event outcomes, i.e., the causal model, the causal model can be further adjusted. For example, in the causal graph, edges that have a relatively small impact on the historical events are "cut". For example, the data analysis device 110 can prune some branches by calculating the weight of the influence element of each node of the causal model. The "cut" process may also be performed by receiving an instruction from a user. For example, factors that have a relatively small impact on the prediction goal are omitted based on the user's experience.

[0077] 5, in block 520, the data analysis device 110 determines at least one goal outcome of interest selected for a current event. The current event can be associated with a historical event. For example, if the historical event is a first quarter promotion situation of a shopping mall, the current event can be, for example, a third quarter promotion situation of the shopping mall.

[0078] In one embodiment, a user instruction for data analysis can be received, and based on the instruction, a user-selected target result of interest for the current event can be determined. The target result can be a possible result of the occurrence of the current event that the user wishes to know. For example, if the event is the shopping mall's third quarter sales promotion status, the target result of interest can be, for example, product sales.

[0079] The instruction may include a number of target results of interest for the user to select, which may be obtained by the occurrence of an event. For example, if the event is the third quarter sales promotion situation of the shopping mall, the target results of interest may be, for example, the increase in product sales, sales profits, and sales volume.

[0080] While the traditional causal model can only select one target outcome, the causal model according to the embodiment of the present disclosure can realize predictions for multiple target outcomes, which can improve the efficiency of data analysis and provide scalability and flexibility to data analysis.

[0081] In block 530, the data analysis device 110 determines current factors influencing the current event and at least one goal object associated with at least one goal result of interest. Again, taking the case where the current event is the third quarter sales promotion situation of the shopping mall as an example, the current factors influencing the current event may include sales promotion methods such as product price reduction, purchase reward system, coupon exchange point system, etc. In one embodiment, the current factors influencing the current event may be obtained by a user's instruction regarding data analysis. In one embodiment, the current factors influencing the current event may be automatically selected from factors that influenced the related historical event based on at least one goal result of interest selected by the user.

[0082] For at least one target object associated with at least one target result of interest, the solution of the present disclosure can divide the target object into different granularities. Figure 7 is a flowchart of a process for determining a target object according to an embodiment of the present disclosure. The determination of the target object will be described below in conjunction with Figure 7.

[0083] In one embodiment, in block 710, the data analysis device 110 can determine a set of objects associated with the data sample of the current condition data of the current factor. Again, taking the current event as an example of the shopping mall's third quarter promotion status, the set of objects can be, for example, the entire consumer population. In block 720, the data analysis device 110 can further determine a desired granularity of grouping.

[0084] In one embodiment, instructions from a user regarding data analysis may be received and the granularity of grouping desired by the user may be determined based on the instructions.

[0085] In some embodiments, the desired granularity of grouping may be, for example, individual target, such as VVIP customers of a shopping mall. If the desired granularity of grouping is individual granularity, a user can obtain training data or a new data file, select one individual by information such as a sample individual ID, and obtain the associated data of the individual as the initial value of the causal model.

[0086] In some embodiments, the desired granularity of grouping may be, for example, a single group of subjects having a predetermined target attribute, such as female consumers aged 28-35. If the desired granularity of grouping is a population granularity, the user may obtain training data or a new data file, and the data analysis device 110 may automatically calculate the average value of each element of the population, and may calculate the initial value of a non-parent node in the causal model, for example, based on the average value of each parent node in the causal graph (see FIG. 2).

[0087] In one embodiment, the desired granularity of grouping may be, for example, multiple groups of objects, each of which has different object attributes. When the desired granularity of grouping is multiple groups, the user obtains training data or a new data file and refines the different populations using refinement criteria, such as a population grouped by gender or a population grouped by product type. Then, a causal model can be established for each of the different groupings, and an initial value for the causal model can be calculated for each of the different groupings.

[0088] At block 730, the data analysis device 110 may determine, based on the grouping granularity, at least one target object from the set of objects that is associated with at least one target outcome of interest.

[0089] In this manner, the causal model according to the embodiments of the present disclosure can achieve more accurate predictions and meet user demands for refined decision making.

[0090] In addition, in one embodiment, the data analysis device 110 can further determine a constraint condition of the current factor. The constraint condition can indicate a selectable data range for the current factor. For example, if the current factor affecting the current event is a product price reduction range, the constraint condition can be, for example, a discount between 5% and 10%. For example, if the current factor affecting the current event is a purchase reward system, the constraint condition can be, for example, a consumption rebate amount that varies from 8% to 10% depending on the product category and equipment. The data analysis device 110 can determine current condition data associated with the current factor based on the constraint condition.

[0091] In one embodiment, instructions from a user regarding data analysis may be received and, based on the instructions, constraints desired by the user may be determined.

[0092] Returning again to FIG. 5, at block 540, the data analysis device 110 determines data for at least one goal outcome of interest based at least on the association, the current condition data associated with the current factor, and the at least one goal object.

[0093] In one embodiment, the association is a causal model as described above. The data analysis device 110 can input the current condition data and at least one target object tag into the causal model, the output of which is the target outcome data.

[0094] In addition, in one embodiment, the data analysis device 110 can obtain condition data samples for at least one target object. The condition data samples can be historical training data samples or new data samples. The data analysis device 110 can determine target outcome data based on the at least one target object based on the condition data samples and the causal model.

[0095] The data analysis solutions described in this disclosure allow for establishing causal models that can be used to evaluate decisions. The models can predict target outcomes for multiple targets and groups of subjects at different granularities. This approach allows for greater flexibility and precision in the analysis, while also meeting user demands for different prediction granularities.

[0096] Exemplary Devices

[0097] FIG. 8 illustrates a block schematic diagram of an exemplary device 800 capable of implementing embodiments of the subject matter of the present disclosure. For example, the data analysis device 110 or the prediction unit 112 illustrated in FIG. 1 can be realized by the device 800. As illustrated, the device 800 includes a central processing unit (CPU) 801. The CPU 801 can perform various appropriate operations and processes based on computer program instructions stored in a read-only memory (ROM) 802 or loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 can further store various programs and data required for the operation of the device 800. The CPU 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) port 805 is also connected to the bus 804.

[0098] A number of components in the device 800 are connected to an I / O port 805. The components include an input unit 806 such as a physical keyboard, a mouse, etc., an output unit 807 such as various types of displays, speakers, etc., a storage unit 808 such as a physical magnetic disk, a physical optical disk, etc., and a communication unit 809 such as a network interface card, a modem, a wireless communication device, etc. The communication unit 809 enables the device 800 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0099] Each of the processes and operations described above, such as methods 500, 600, and 700, may be executed by the processing unit 801. For example, in some embodiments, the methods 500, 600, and 700 may be implemented as a computer software program and tangibly stored in a device readable medium, such as the storage unit 808. In some embodiments, some or all of the computer program may be loaded and / or installed in the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded into the RAM 803 and executed by the CPU 801, it may perform one or more operations of the methods 500, 600, and 700 described above.

[0100] The present disclosure may be a method, an apparatus, a system, and / or a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions stored thereon for carrying out aspects of the present disclosure.

[0101] A computer readable storage medium may be a tangible device capable of holding and storing instructions for use by an instruction execution device. A computer readable storage medium may be, for example, but not limited to, an electrical storage unit, a magnetic storage unit, an optical storage unit, an electromagnetic storage unit, a semiconductor storage unit, or any suitable combination thereof. More specific examples (but not all) of computer readable storage media include a physical portable computer diskette, a physical hard disk, a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact physical disk read only memory (CD-ROM), a digital versatile physical disk (DVD), a memory stick, a physical floppy disk, a mechanical encoder disk, a punch card or a protruding structure in a groove on which instructions are stored, and any suitable combination thereof. A computer-readable storage medium as used herein is not to be understood as being a momentary signal per se, such as, for example, radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through wave guides or other transmission media (e.g., light pulses through optical cables), or electrical signals transmitted over electrical wires.

[0102] The computer readable program instructions described herein can be downloaded from a computer readable storage medium to each computing / processing device, or can be downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical transmission cables, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. A network interface card or network interface in each computing / processing device receives the computer readable program instructions from the network and forwards the computer readable program instructions to be stored in the computer readable storage medium of each computing / processing device.

[0103] The computer program instructions for carrying out the operations of the present disclosure may be assembler directives, instruction set architecture (ISA), machine language instructions, machine related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and general process-based programming languages ​​such as "C" or similar programming languages. The computer readable program instructions may be entirely executed on the user computer, partially executed on the user computer, executed as a separate software package, partially executed on the user computer and partially executed on a remote computer, or entirely executed on a remote computer or server. In the context of a remote computer, the remote computer may be connected to the user computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet Service Provider). In some embodiments, the status information of the computer readable program instructions can be used to customize electronic circuitry, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), that can execute the computer readable program instructions to implement aspects of the present disclosure.

[0104] Aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0105] These computer readable program instructions may be provided to a processor unit of a general purpose computer, a special purpose computer, or other programmable data processing apparatus to generate an apparatus that, when executed by the processor unit of the computer or other programmable data processing apparatus, generates an apparatus that implements the functions / operations defined in one or more blocks of the flowcharts and / or block diagrams. These computer readable program instructions may be stored in a computer readable storage medium. These instructions cause a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, a computer readable medium having instructions stored thereon includes an article of manufacture that includes instructions for each aspect of implementing the functions / operations defined in one or more blocks of the flowcharts and / or block diagrams.

[0106] The computer readable program instructions may be loaded into a computer, other programmable data processing apparatus, or other device to cause the computer, other programmable data processing apparatus, or other device to perform a series of operational steps to generate a computer implemented process, such that the instructions executing on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams.

[0107] The flowcharts and block diagrams in the figures represent possible architectures, functions and operations of systems, methods and computer program products according to embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams may represent a module, program segment or part of instructions, which includes one or more executable instructions for implementing a specified logical function. Alternatively, in some implementations, the functions depicted in the blocks may occur in a different order than depicted in the figures. For example, two consecutive blocks may actually be essentially executed in parallel, but may be executed in the opposite order in some cases. This is determined by the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented in a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0108] Although each embodiment of the present disclosure has been described above, the above description is illustrative and not exhaustive, and is not limited to each of the disclosed embodiments. It is clear that a person skilled in the art can make multiple modifications and changes without departing from the scope and spirit of each of the described embodiments. The terms used herein are selected with the intention of optimally explaining the principles, actual applications, or technical improvements in the market of each embodiment, or allowing a person skilled in the art to understand each embodiment disclosed in the present specification.

Claims

1. A data analysis device, generating an association between historical factors and the outcome of the historical event based at least on historical condition data associated with historical factors that influenced the historical event and historical outcome data indicative of an outcome of the historical event caused by the historical factors; determining at least one goal outcome of interest selected for a current event associated with the historical event; determining current factors influencing the current event and at least one goal object associated with the at least one goal outcome of interest; determining data for the at least one goal outcome of interest based at least on the association, current condition data associated with the current factor, and the at least one goal object; The method used for data analysis, including:

2. The data analysis device further includes acquiring historical condition data, The historical condition data includes at least one of continuous historical condition data and discrete historical condition data. The method of claim 1.

3. The acquiring of the historical condition data includes: periodically obtaining said historical condition data from a data collection device. The method of claim 2.

4. The generating of the association comprises: acquiring reference data indicating a degree of influence of the history condition data on the history result data in response to the number of data samples of the acquired history condition data being less than a threshold number; determining the historical result data based on the historical condition data and the reference data; determining the association based on the historical condition data and the historical result data; The method of claim 2 , comprising:

5. Determining the at least one target outcome of interest includes: receiving a first instruction regarding the data analysis from a user terminal; determining at least one target outcome of interest based on the first indication; The method of claim 1 , comprising:

6. The data analysis device determines a constraint for the current factor indicating a selectable data range for the current factor; and the data analysis device determining the current condition data associated with the current factor based on the constraint; The method of claim 1 further comprising:

7. Determining the constraints includes: obtaining second instructions regarding the data analysis from a user terminal; determining the constraint condition based on the second instruction; and The method of claim 6, comprising:

8. determining the target object comprises: determining a set of objects associated with the data sample of current condition data of the current factor; Determining a desired grouping granularity; determining, from among the set of objects, the at least one target object that is associated with the at least one target outcome of interest based on the granularity of the grouping; The method of claim 1 , comprising:

9. Determining the granularity of the grouping comprises: obtaining a third instruction regarding the data analysis from a user terminal; determining a granularity of the grouping based on the third instruction; The method of claim 8 , comprising:

10. The granularity of the grouping is: Individual subjects, a single group of subjects having predetermined subject attributes; and Multiple groups of subjects with different target attributes At least one of: The method according to claim 8.

11. the association is a causal model; Determining the data of the target outcome includes: inputting the current condition data and a tag of the at least one target object into the causal model to determine the data of the target outcome. The method of claim 1.

12. The data analysis device further includes outputting the data of the determined target result to a user terminal. The method of claim 1.

13. The data analysis device further comprises displaying the data of the determined target result on a user terminal. The method of claim 1.

14. The method of claim 13, further comprising: obtaining a predetermined reference amount of the target result; In response to the data analysis device determining that a difference between the data and the predetermined reference amount exceeds a threshold difference, outputting a signal representative of the difference to a user terminal. The method of claim 1 further comprising:

15. A device for use in data analysis, comprising: At least one processor; a memory coupled to said at least one processor, storing instructions which, when executed by said at least one processor, cause said device to perform a method according to any one of claims 1 to 14; Including, the device.

16. A computer program which, when executed by a processor, causes a computer to carry out the method according to any one of claims 1 to 14. Computer program.

Images