Idle period prediction method and system and preset machine learning model training method
By combining the behavioral data of the target service object and external environment factors, using machine learning models to predict the first idle period, and secondary prediction through attribute information and statistical models, the problem of low idle period prediction in the existing technology is solved, achieving higher prediction accuracy.
Patent Information
- Application Number
- CN202411711928.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2025-03-04
AI Technical Summary
In the prior art, when making idle period prediction, the prediction accuracy is low.
By obtaining the behavioral data of the target service object and external environmental impact factors, the first idle period prediction is made using the preset machine learning model to obtain the first predicted idle period at the hour level. Then, based on the attribute information of the target service object and the first predicted idle period, a secondary idle period prediction is performed using a preset statistical model to obtain a second predicted idle period of half an hour.
Through the dual prediction method, the accuracy of idle time prediction is improved, and the problem of low accuracy caused by the sparseness of hourly prediction data is solved.
Smart Images
Figure CN119558900A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to an idle time period prediction method, system and preset machine learning model training method. Background Art
[0002] In traditional telemarketing scenarios, telemarketing is often conducted based on fixed time periods. Since different target service objects have different special circumstances, telemarketing based on fixed time periods will bring poor experience to the target service objects. In related technologies, the idle time periods of target service objects are usually predicted based on information such as holidays, but this method has the problem of low prediction accuracy.
[0003] Therefore, how to solve the problem of low prediction accuracy when related technologies predict idle time periods has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention
[0004] Based on the above problems, in order to improve the prediction accuracy for idle time periods, the embodiments of the present application provide an idle time period prediction method, system and preset machine learning model training method.
[0005] The embodiments of the present application disclose the following technical solutions:
[0006] In a first aspect, an embodiment of the present application provides an idle period prediction method, comprising:
[0007] Obtain behavioral data of target service recipients and external environmental influencing factors;
[0008] Performing a first idle period prediction based on the behavior data, the external environment influencing factors, and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is hourly level;
[0009] Acquire the attribute information of the target service object, and perform a secondary idle period prediction on the target service object based on the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
[0010] In a possible implementation, performing a secondary idle period prediction on the target service object according to the attribute information, the first predicted idle period, and a preset statistical model to obtain a second predicted idle period includes:
[0011] Performing customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object;
[0012] The target preset customer group and the first predicted idle time period are used as input data of the preset statistical model to determine the second predicted idle time period.
[0013] In a possible implementation, the method further includes:
[0014] Obtaining a third predicted idle time period corresponding to each of the plurality of service objects; the predicted idle time period is predicted by the preset machine learning model and has a coarse granularity of hourly level;
[0015] Classify customer groups according to the attribute information of each service object to obtain multiple preset customer groups; the preset customer groups include multiple service objects;
[0016] The plurality of third predicted idle time periods and the plurality of preset customer groups are used as input data of the preset statistical model to establish a mapping relationship table between the plurality of preset customer groups and the plurality of fourth predicted idle time periods; the coarse granularity of the fourth predicted idle time period is half an hour level.
[0017] In a possible implementation, after acquiring the attribute information of the target service object and performing a secondary idle period prediction on the target service object according to the attribute information, the first predicted idle period and a preset statistical model to obtain the second predicted idle period, the method further includes:
[0018] When there are multiple second predicted idle time periods, score the idleness of each second predicted idle time period by using a multi-label model to obtain an idleness score of each second predicted idle time period;
[0019] Marketing resources for the target service object are allocated according to the idleness score of each of the second predicted idle time periods.
[0020] In a second aspect, an embodiment of the present application provides a preset machine learning model training method, comprising:
[0021] Obtain behavioral data and external environmental influencing factors of multiple service objects;
[0022] Performing feature processing on the behavior data of each service object and the external environmental factors to obtain a free time period prediction feature set;
[0023] A machine learning model is trained based on the idle period prediction feature set to obtain the preset machine learning model.
[0024] In a possible implementation, the behavior data includes: historical interaction records and historical interaction data;
[0025] The feature processing of the behavior data of each service object and the external environmental factors to obtain a free time prediction feature set includes:
[0026] Perform frequency analysis based on the historical interaction records to obtain multiple historical interaction time periods;
[0027] Processing the historical interaction data in each of the historical interaction time periods to obtain an interaction success rate in each of the historical interaction time periods;
[0028] The interaction success rate and the external environment influencing factor in each of the historical interaction time periods are subjected to feature fusion processing to obtain the idle period prediction feature set.
[0029] In one possible implementation, the preset machine learning model is used to implement any possible idle period prediction method as described in the first aspect.
[0030] In a third aspect, an embodiment of the present application provides an idle period prediction system, including:
[0031] An acquisition module is used to acquire the behavior data of the target service object and the external environment influencing factors;
[0032] A first prediction module is used to predict the first idle period according to the behavior data, external environment influencing factors and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is hourly level;
[0033] The second prediction module is used to obtain the attribute information of the target service object, and perform a secondary idle period prediction on the target service object based on the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
[0034] In a possible implementation, the second prediction module is specifically configured to:
[0035] Performing customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object;
[0036] The target preset customer group and the first predicted idle time period are used as input data of the preset statistical model to determine the second predicted idle time period.
[0037] In a possible implementation, the system further includes: a mapping relationship table establishment module; the mapping relationship table establishment module is specifically used to:
[0038] Obtaining third predicted idle time periods corresponding to each of the plurality of service objects; the predicted idle time periods are predicted by the preset machine learning model; the coarse granularity of the third predicted idle time periods is hourly level;
[0039] Classify customer groups according to the attribute information of each service object to obtain multiple preset customer groups; the preset customer groups include multiple service objects;
[0040] The plurality of third predicted idle time periods and the plurality of preset customer groups are used as input data of the preset statistical model to establish a mapping relationship table between the plurality of preset customer groups and the plurality of fourth predicted idle time periods; the coarse granularity of the fourth predicted idle time period is half an hour level.
[0041] Compared with the prior art, the present application has the following beneficial effects: the embodiment of the present application provides an idle time prediction method, system and preset machine learning model training method. In the idle time prediction method, the behavior data and external environmental factors of the target service object are first obtained, and the first idle time prediction is performed by the preset machine learning model and the obtained data to obtain the first predicted idle time at the hour level. Among them, by combining the behavior data of the target service object and the actual influencing factors of the external environment, the idle time prediction can be performed in all aspects to obtain the first predicted idle time. On the other hand, since the idle time prediction data at the hour level is often sparse and the data accuracy is low, in order to further improve the prediction accuracy of the idle time, on the basis of obtaining the first predicted idle time, a preset statistical model is used, combined with the attribute information of the target service object and the previously determined first predicted idle time, a second idle time prediction is performed to obtain a second predicted time at the half-hour level, so as to deal with the problem of sparsity of data in the first predicted idle time at the hour level. In this way, by predicting the double idle time of the target service object, the predicted idle time at the half-hour level is determined from many aspects, thereby comprehensively improving the prediction accuracy of the idle time. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0043] Figure 1 A flowchart of an idle period prediction method provided in an embodiment of the present application;
[0044] Figure 2A flowchart of a preset machine learning model training method provided in an embodiment of the present application;
[0045] Figure 3 A flowchart of another idle period prediction method provided in an embodiment of the present application;
[0046] Figure 4 A schematic diagram of the structure of an idle period prediction system provided in an embodiment of the present application. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solutions and advantages of this application more clear, the following is a further detailed description of this application in combination with specific embodiments and with reference to the accompanying drawings. It should be noted that the embodiments described in the embodiments of this application are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0048] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "include" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right", etc. are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0049] As described above, in traditional telemarketing scenarios, telemarketing is often conducted based on fixed time periods. Since different target service objects have different special circumstances, telemarketing based on fixed time periods will bring poor experience to the target service objects. In related technologies, the idle time periods of target service objects are usually predicted based on information such as holidays, but this method has the problem of low prediction accuracy.
[0050] In order to solve the above problems, the embodiment of the present application provides an idle time prediction method, system and preset machine learning model training method. In the idle time prediction method, the behavior data and external environmental factors of the target service object are first obtained, and the first idle time prediction is performed by the preset machine learning model and the obtained data to obtain the first predicted idle time at the hour level. Among them, by combining the behavior data of the target service object and the actual influencing factors of the external environment, the idle time prediction can be performed in all aspects to obtain the first predicted idle time. On the other hand, since the idle time prediction data at the hour level is often sparse and the data accuracy is low, in order to further improve the prediction accuracy of the idle time, on the basis of obtaining the first predicted idle time, a preset statistical model is used, combined with the attribute information of the target service object and the previously determined first predicted idle time, a second idle time prediction is performed to obtain a second predicted time at the half-hour level, so as to deal with the problem of sparsity of data in the first predicted idle time at the hour level. In this way, by predicting the double idle time of the target service object, the predicted idle time at the half-hour level is determined from many aspects, thereby comprehensively improving the prediction accuracy of the idle time.
[0051] 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 are within the scope of protection of this application.
[0052] It should be noted that the behavioral data (including but not limited to the device information of the target service object, the attribute information of the target service object, the behavioral data of the target service object, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the target service object or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.
[0053] Next, the idle time period prediction method provided by the embodiment of the present application will be introduced in conjunction with the specific embodiment drawings. Figure 1 , which is a flow chart of an idle period prediction method provided in an embodiment of the present application, and specifically includes the following steps:
[0054] S101: Obtaining the behavior data of the target service object and external environment influencing factors.
[0055] First, in order to accurately verify the idle time of the target service object, it is necessary to obtain the behavioral data of the target service object and the external environment impression factors from multiple channels.
[0056] Among them, the behavior data of the target service object may be historical call data, online behavior data, offline behavior data, and the like. Furthermore, the historical call data may be data such as the number of calls, call duration, and call frequency between the target service object and the service provider in different historical time periods. Online behavior data may be data such as its active status on the service provider's APP, and execution operations. Correspondingly, offline behavior data may be the interaction between the target service object and the service provider offline, and the like. The core of acquiring the behavior data of the target service object is to understand the interaction between the target service object and the service provider in different historical time periods, so as to predict the idle time period of the target service object.
[0057] External environmental influencing factors are used to characterize the impact of external objective environmental factors on idle time, such as holidays, major events, and lunch breaks, etc. By combining the behavioral data of the target service object and the external environmental influencing factors, idle time prediction can be performed based on the actual situation of the target service object on the basis of the reference to the actual situation of the external environment, thereby ensuring the accuracy of the prediction.
[0058] S102: Perform an initial idle period prediction based on the behavior data, the external environment influencing factors, and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is at the hour level.
[0059] Furthermore, the target service object is predicted for the first idle time period through the pre-trained preset machine learning model and the real-time acquired data, so as to obtain the first predicted idle time period at the coarse-grained hour level. In this process, the behavior data of the target service object and the external environmental influencing factors are used as the input data of the preset machine learning model. The processing characteristics of the machine learning model are used to comprehensively consider the multi-dimensional behavior data and external environmental influencing factors. By learning the complex relationship between these features, the idle time period is predicted according to the personal behavior pattern and preference of each target service object, rather than adopting a one-size-fits-all prediction method, so as to improve the accuracy of the idle time period prediction.
[0060] In this step, the coarse-grainedness of the first predicted idle time period is hourly, such as (14:00-15:00), (9:00-10:00), etc. The hourly predicted idle time period can preliminarily characterize the possible idle time period of the target service object, but the data has a certain data sparsity. If the first predicted idle time period with a coarse-grained hourly level is used as the final idle time period prediction result for the target service object, there may be a problem of low idle time period prediction accuracy. Therefore, this step is only used as the first prediction of the overall idle time period prediction. In the subsequent steps, it is necessary to perform a secondary idle time period prediction to reduce the coarse-grainedness of the idle time period prediction result, thereby improving the idle time period prediction accuracy.
[0061] In actual application scenarios, the model type of the preset machine learning model can be selected from a variety of different types according to needs, such as decision trees, random forests, and neural networks. The machine learning model can help capture the nonlinear relationship between features, thereby effectively improving the accuracy of idle period prediction in the embodiment of the present application.
[0062] Next, the training method of the preset machine learning model in the embodiment of the present application will be introduced in conjunction with the specific embodiment drawings. Figure 2 , which is a flow chart of a preset machine learning model training method provided in an embodiment of the present application, and specifically includes the following steps:
[0063] S201: Obtaining behavior data and external environment influencing factors of multiple service objects.
[0064] Similar to the above-mentioned idle time prediction method for the target service object, in the training stage of the preset machine learning model, it is necessary to obtain the behavior data of multiple service objects themselves and their corresponding external environmental influencing factors, and use them as training data for the preset machine learning model.
[0065] In a possible implementation, the external environment influencing factor may correspond to the city and region where each service object is located. For example, when the external environment influencing factor is a major event, different cities may have their own special festivals. Therefore, the external environment influencing factor may also be determined based on the city and region information of the service object, which is not described in detail in this embodiment.
[0066] S202: Perform feature processing on the behavior data of each of the service objects and the external environmental factors to obtain a free time prediction feature set.
[0067] Subsequently, due to the large differences in data structure between the original behavior data of the service object and the external environmental factors, it is necessary to perform feature processing on both in order to obtain the idle time prediction feature set for training the preset machine learning model. Specifically, the process of feature processing the behavior data and external environmental influencing factors is achieved through the following three steps:
[0068] Step 1: Perform frequency analysis based on the historical interaction records to obtain multiple historical interaction time periods.
[0069] The behavioral data of the service object includes historical interaction records and historical interaction data between the service object and the service provider during the historical period. These historical interaction records may be historical call records or offline activity records between the service object and the service provider, and the corresponding historical interaction data may also be online call content, offline interaction content, specific interaction time, etc.
[0070] By performing frequency analysis on the service object's historical interaction records, it is possible to determine the time periods in which the service object and the service provider have a higher interaction frequency, thereby determining the historical interaction time periods and using them as the data basis for predicting idle time periods for the service object.
[0071] Step 2: Process the historical interaction data in each of the historical interaction time periods to obtain the interaction success rate in each of the historical interaction time periods.
[0072] After determining the historical interaction time period of the service object, data processing is performed through the historical interaction data in each historical interaction time period, such as calculating the number of successful interactions (number of phone calls connected, number of valid replies, etc.) and the total number of interactions in each historical interaction time period. Based on the number of successful interactions and the total number of interactions, the interaction success rate in each historical interaction time period can be calculated.
[0073] Step three: perform feature fusion processing on the interaction success rate and the external environment influencing factor in each of the historical interaction time periods to obtain the idle time period prediction feature set.
[0074] Finally, in the process of feature fusion processing, it is necessary to merge the interaction success rate and external environment influencing factors in each historical interaction time period according to the ID of the service object and the historical interaction time period, so as to form a new idle period prediction feature set.
[0075] S203: Perform machine learning model training based on the idle period prediction feature set to obtain the preset machine learning model.
[0076] In this way, necessary data processing is performed on the merged feature set, such as normalization and missing data processing, to ensure that the feature set can meet the input requirements of the preset machine learning model.
[0077] The above is an introduction to the preset machine learning model training method. Next, we will continue to combine Figure 1 The idle time period prediction method provided in the embodiment of the present application is introduced as follows:
[0078] S103: Acquire the attribute information of the target service object, and perform a secondary idle period prediction on the target service object according to the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
[0079] From the above description of the first predicted idle period, it can be seen that the first predicted idle period with a coarse granularity of hourly level has a certain degree of data sparsity problem, and the first predicted idle period obtained through the first idle period prediction cannot be used as the final idle period prediction result.
[0080] Therefore, in order to improve the accuracy of the idle time prediction for the target service object, on the basis of the first prediction based on its behavioral data and external environmental influencing factors, it is necessary to further perform a secondary idle time prediction based on the attribute information of the target service object. The attribute information of the target service object is more used to characterize the specific information of the target service object itself, such as age, city, gender, etc., combined with the attribute information of the target service object, the customer group category of the target service object can be determined, so that the idle time prediction result for the entire customer group category can be used as a reference factor to improve the accuracy of the target service object prediction. Therefore, it is necessary to use a preset statistical model to perform a secondary idle time prediction to obtain the second predicted idle time, that is, the final idle time prediction result for the target service object.
[0081] Next, the method for performing secondary idle time period prediction based on the apartment type information of the target service object and the preset statistical model in step S103 will be introduced in conjunction with the drawings of the specific embodiments.
[0082] See also Figure 3 , which is a flow chart of another idle period prediction method provided in an embodiment of the present application, which specifically includes the following steps:
[0083] S1031: Perform customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object.
[0084] First, according to the attribute information of the target service object, the target service object is analyzed for customer groups to determine the target preset customer groups corresponding to the target service object. The preset customer groups are customer group classifications pre-constructed in the embodiment of the present application, which are classified by various attribute information such as age, gender, occupation, etc.
[0085] In an embodiment of the present application, an unsupervised clustering algorithm such as KMeans or hierarchical clustering can be used to establish a customer clustering model, and this can be used to combine the attribute information of each service object to classify the customer groups, and different labels are assigned to each different customer group to obtain multiple preset customer groups.
[0086] S1032: Using the target preset customer group and the first predicted idle time period as input data of the preset statistical model to determine the second predicted idle time period.
[0087] Furthermore, the target preset customer group to which the target service object belongs and the first predicted idle time period previously determined are used as input data of the preset statistical model. The preset statistical model can count the half-hour time period in which each preset customer group appears most in each hour-level coarse-grained time period through the time periods divided at the one-hour level in the first predicted idle time period. In this way, through the target preset customer group corresponding to the target service object itself and the first predicted idle time period, the second predicted idle time period at the half-hour level can be further determined to improve the accuracy of idle time period prediction for the target service object.
[0088] For example, the first predicted idle time period of the target service object is set to (14:00-15:00), and its corresponding preset customer group is customer group A. Through the preset statistical model and multiple pre-classified preset customer groups, it is determined that the predicted idle time period included in customer group A is (14:30-15:00). In this case, the first predicted idle time period of the target service object (14:00-15:00) and the predicted idle time period corresponding to customer group A (14:30-15:00) can be intersected to obtain the value, so that the final idle time period prediction result of the target service object can be determined to be (14:30-15:00), that is, the second predicted idle time period of the target service object is (14:30-15:00).
[0089] In a possible implementation, if multiple half-hour second predicted idle time periods are finally predicted, a multi-label model can be further used to score the idleness of each second predicted idle time period to determine the idleness score corresponding to each second predicted idle time period. When providing services according to the second predicted idle time period of the target service object, the marketing resources corresponding to the service are allocated according to the idleness score corresponding to each time period, thereby maximizing the marketing efficiency.
[0090] Specifically, the idleness score of each second predicted idle time period can be scored by the call connection rate corresponding to the idle time period. The higher the call connection rate, the higher the corresponding idleness score. This embodiment of the present application will not be elaborated on this.
[0091] On the other hand, from the description of the preset customer groups in the previous text, it can be known that multiple service objects can be divided into multiple preset customer groups through the attribute information of each service object. In this case, through the third predicted idle time period at the coarse-grained hour level corresponding to each service object and its corresponding preset customer group, it can be used as the input data of the preset statistical model, on the basis of establishing the customer group to which each service object belongs, according to the third predicted idle time period of each service object, the half-hour time period with the most appearance of each customer group can be counted, and the corresponding mapping relationship table can be output.
[0092] In this way, a mapping relationship table between multiple preset customer groups and multiple fourth predicted idle time periods with a coarse-grained level of half an hour can be established, that is, the predicted idle time period with a coarse-grained level of half an hour corresponding to each customer group is determined. When the idle time period prediction is subsequently performed for a certain target service object, the idle time period prediction for a single target service object can be completed through the mapping relationship table of "service object-preset customer group-fourth predicted idle time period", thereby improving the prediction efficiency for the idle time period.
[0093] The embodiment of the present application provides an idle time prediction method and a preset machine learning model training method. In the idle time prediction method, the behavior data and external environmental factors of the target service object are first obtained, and the first idle time prediction is performed through the preset machine learning model and the obtained data to obtain the first predicted idle time at the hour level. Among them, by combining the behavior data of the target service object and the actual influencing factors of the external environment, the idle time prediction can be performed in all aspects to obtain the first predicted idle time. On the other hand, since the idle time prediction data at the hour level is often sparse and the data accuracy is low, in order to further improve the prediction accuracy of the idle time, on the basis of obtaining the first predicted idle time, a preset statistical model is used, combined with the attribute information of the target service object and the previously determined first predicted idle time, a second idle time prediction is performed to obtain a second predicted time at the half-hour level, so as to deal with the problem of sparsity of data in the first predicted idle time at the hour level. In this way, by predicting the double idle time of the target service object, the predicted idle time at the half-hour level is determined from many aspects, thereby comprehensively improving the prediction accuracy of the idle time.
[0094] An idle time period prediction system provided in an embodiment of the present application is introduced below. The idle time period prediction system described below and the idle time period prediction method described above can refer to each other.
[0095] See also Figure 4 , which is a schematic diagram of the structure of an idle time period prediction system provided in an embodiment of the present application, specifically including the following modules:
[0096] An acquisition module 100 is used to acquire the behavior data of a target service object and external environment influencing factors;
[0097] A first prediction module 200 is used to predict the first idle period according to the behavior data, external environmental influencing factors and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is hourly level;
[0098] The second prediction module 300 is used to obtain the attribute information of the target service object, and perform a secondary idle period prediction on the target service object based on the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
[0099] In a possible implementation, the second prediction module 300 is specifically configured to:
[0100] Performing customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object;
[0101] The target preset customer group and the first predicted idle time period are used as input data of the preset statistical model to determine the second predicted idle time period.
[0102] In a possible implementation, the system further includes: a mapping relationship table establishment module; the mapping relationship table establishment module is specifically used to:
[0103] Obtaining third predicted idle time periods corresponding to each of the plurality of service objects; the predicted idle time periods are predicted by the preset machine learning model; the coarse granularity of the third predicted idle time periods is hourly level;
[0104] Classify customer groups according to the attribute information of each service object to obtain multiple preset customer groups; the preset customer groups include multiple service objects;
[0105] The plurality of third predicted idle time periods and the plurality of preset customer groups are used as input data of the preset statistical model to establish a mapping relationship table between the plurality of preset customer groups and the plurality of fourth predicted idle time periods; the coarse granularity of the fourth predicted idle time period is half an hour level.
[0106] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, an embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the idle period prediction method described in any of the above embodiments.
[0107] The computer-readable media of the embodiments of the present application include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.
[0108] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the idle period prediction method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0109] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the method and system, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The method and system described above are merely schematic, in which the units described as separate components may or may not be physically separated, and the components indicated as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.
[0110] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for predicting idle time periods, characterized in that: include: Obtain behavioral data of target service recipients and external environmental influencing factors; Performing a first idle period prediction based on the behavior data, the external environment influencing factors, and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is hourly level; Acquire the attribute information of the target service object, and perform a secondary idle period prediction on the target service object based on the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
2. The method according to claim 1, characterized in that: The performing a secondary idle period prediction on the target service object according to the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period includes: Performing customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object; The target preset customer group and the first predicted idle time period are used as input data of the preset statistical model to determine the second predicted idle time period.
3. The method according to claim 1, characterized in that The method further comprises: Obtaining a third predicted idle time period corresponding to each of the plurality of service objects; the predicted idle time period is predicted by the preset machine learning model and has a coarse granularity of hourly level; Classify customer groups according to the attribute information of each service object to obtain multiple preset customer groups; the preset customer groups include multiple service objects; The plurality of third predicted idle time periods and the plurality of preset customer groups are used as input data of the preset statistical model to establish a mapping relationship table between the plurality of preset customer groups and the plurality of fourth predicted idle time periods; the coarse granularity of the fourth predicted idle time period is half an hour level.
4. The method according to claim 1, characterized in that After acquiring the attribute information of the target service object and performing a secondary idle period prediction on the target service object according to the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period, the method further includes: When there are multiple second predicted idle time periods, score the idleness of each second predicted idle time period by using a multi-label model to obtain an idleness score of each second predicted idle time period; Marketing resources for the target service object are allocated according to the idleness score of each of the second predicted idle time periods.
5. A preset machine learning model training method, characterized in that: include: Obtain behavioral data and external environmental influencing factors of multiple service objects; Performing feature processing on the behavior data of each service object and the external environmental factors to obtain a free time period prediction feature set; A machine learning model is trained based on the idle period prediction feature set to obtain the preset machine learning model.
6. The method according to claim 5, characterized in that The behavior data includes: historical interaction records and historical interaction data; The feature processing of the behavior data of each service object and the external environmental factors to obtain a free time prediction feature set includes: Perform frequency analysis based on the historical interaction records to obtain multiple historical interaction time periods; Processing the historical interaction data in each of the historical interaction time periods to obtain an interaction success rate in each of the historical interaction time periods; The interaction success rate and the external environment influencing factor in each of the historical interaction time periods are subjected to feature fusion processing to obtain the idle period prediction feature set.
7. The method according to claim 5, characterized in that The preset machine learning model is used to implement the idle period prediction method as described in any one of claims 1-4.
8. A system for predicting idle time periods, characterized in that: include: An acquisition module is used to acquire the behavior data of the target service object and the external environment influencing factors; A first prediction module is used to predict the first idle period according to the behavior data, external environment influencing factors and a preset machine learning model to obtain a first predicted idle period; the coarse granularity of the first predicted idle period is hourly level; The second prediction module is used to obtain the attribute information of the target service object, and perform a secondary idle period prediction on the target service object based on the attribute information, the first predicted idle period and a preset statistical model to obtain a second predicted idle period; the coarse granularity of the second predicted idle period is half an hour.
9. The system according to claim 8, characterized in that The second prediction module is specifically used for: Performing customer group analysis on the target service object based on the attribute information to obtain a target preset customer group corresponding to the target service object; The target preset customer group and the first predicted idle time period are used as input data of the preset statistical model to determine the second predicted idle time period.
10. The system according to claim 8, characterized in that Also includes: A mapping relationship table establishment module; the mapping relationship table establishment module is specifically used to: Obtaining third predicted idle time periods corresponding to each of the plurality of service objects; the predicted idle time periods are predicted by the preset machine learning model; the coarse granularity of the third predicted idle time periods is hourly level; Classify customer groups according to the attribute information of each service object to obtain multiple preset customer groups; The preset customer group includes a plurality of the service objects; Using the plurality of third predicted idle time periods and the plurality of preset customer groups as input data of the preset statistical model to establish a mapping relationship table between the plurality of preset customer groups and the plurality of fourth predicted idle time periods; The coarse granularity of the fourth predicted idle period is half an hour level.