Strategy recommendation method, device, equipment and storage medium

Abstract

The application discloses a strategy recommendation method and device, equipment and a storage medium. The method comprises the following steps: obtaining a training data set, training an initial prediction model based on the training data set, and obtaining a target prediction model; obtaining a to-be-recommended strategy and inputting the to-be-recommended strategy into the target prediction model; the target prediction model outputs an acceptance probability of the to-be-recommended strategy based on a historical probability of a historical state of a previous historical recommended strategy of the to-be-recommended strategy; comparing the acceptance probability with an acceptance probability threshold to obtain a comparison result; and determining the to-be-recommended strategy as a target recommended strategy according to the comparison result, and recommending the target recommended strategy. The technical scheme of the application predicts the acceptance probability of the to-be-recommended strategy accepted by a user, and when the acceptance probability of the to-be-recommended strategy is relatively small, the to-be-recommended strategy is not recommended to the user, the influence of strategy recommendation on the user is reduced, and the satisfaction of the user is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of artificial intelligence and big data, and particularly relates to a strategy recommendation method and device, equipment and a storage medium. BACKGROUND

[0002] In a points-of-interest (POI) recommendation scenario, the existing recommendation scenario performs recommendation once a certain specific trigger condition is met; however, such recommendation is uncertain to be accepted by a user, and many recommendations may be rejected by the user; therefore, such recommendation mode may affect the normal performance of a continuous operation currently performed by the user, and further reduce user satisfaction. SUMMARY

[0003] The present application aims to at least solve one of the technical problems in the related art. To this end, one object of the present application is to provide a strategy recommendation method and device, equipment and a storage medium.

[0004] To solve the above technical problems, embodiments of the present application provide the following technical solutions:

[0005] A strategy recommendation method comprises:

[0006] obtaining a training data set, and training an initial prediction model based on the training data set to obtain a target prediction model;

[0007] obtaining a to-be-recommended strategy, and inputting the to-be-recommended strategy into the target prediction model;

[0008] The target prediction model outputs an acceptance probability of the to-be-recommended strategy based on a historical probability of a historical state of a previous historical recommended strategy of the to-be-recommended strategy;

[0009] comparing the acceptance probability with an acceptance probability threshold to obtain a comparison result;

[0010] According to the comparison result, the to-be-recommended strategy is determined as a target recommended strategy, and the target recommended strategy is recommended.

[0011] Optionally, the obtaining a training data set, and training an initial prediction model based on the training data set to obtain a target prediction model comprises:

[0012] classifying the training data set based on a preset rule to obtain a plurality of historical data sets; wherein each historical data set comprises a plurality of historical data; each historical data comprises a historical recommended strategy, a historical state corresponding to the historical recommended strategy, a historical probability of the historical state, and a time stamp;

[0013] obtaining the timestamp of each of the historical data in each of the historical data sets;

[0014] sorting the historical data in each of the historical data sets based on the timestamp, to obtain a plurality of historical time-series data sets;

[0015] training the initial prediction model based on the plurality of historical time-series data sets, to obtain the target prediction model.

[0016] Optionally, the training of the initial prediction model based on the plurality of historical time-series data sets, to obtain the target prediction model, comprises:

[0017] obtaining a historical to-be-recommended strategy;

[0018] determining a target historical time-series data set according to the historical to-be-recommended strategy;

[0019] obtaining a target historical state chain based on a plurality of historical states in the target historical time-series data set; wherein the historical state comprises a historical acceptance state or a historical non-acceptance state;

[0020] calculating the acceptance probability of the historical to-be-recommended strategy based on the target historical state chain.

[0021] Optionally, the calculating of the acceptance probability of the historical to-be-recommended strategy based on the target historical state chain comprises:

[0022] calculating a first numerical value according to the target historical state chain; wherein the first numerical value represents the number of the historical acceptance states included in the target historical state chain;

[0023] calculating a second numerical value according to the target historical state chain; wherein the second numerical value represents the number of the historical non-acceptance states included in the target historical state chain;

[0024] calculating a target matrix of the target historical state chain according to the target historical state chain, the first numerical value and the second numerical value;

[0025] calculating the acceptance probability of the historical to-be-recommended strategy according to the target matrix of the target historical state chain and the historical probability of the historical state of a previous historical recommended strategy of the historical to-be-recommended strategy.

[0026] Optionally, the calculating of the target matrix of the target historical state chain according to the target historical state chain, the first numerical value and the second numerical value comprises:

[0027] a first transition value is calculated, wherein the first transition value is used to represent a number of transitions from the historical accepted state to the historical accepted state in the target historical state chain;

[0028] a second transition value is calculated, wherein the second transition value is used to represent a number of transitions from the historical accepted state to the historical unaccepted state in the target historical state chain;

[0029] a third transition value is calculated, wherein the third transition value is used to represent a number of transitions from the historical unaccepted state to the historical accepted state in the target historical state chain;

[0030] a fourth transition value is calculated, wherein the fourth transition value is used to represent a number of transitions from the historical unaccepted state to the historical unaccepted state in the target historical state chain;

[0031] The target matrix of the target historical state chain is calculated based on the first transition value, the second transition value, the third transition value, and the fourth transition value.

[0032] Optionally, the obtaining of the historical probability of the historical state of the historical recommended strategy comprises:

[0033] The timestamp T of the historical recommended strategy is obtained.

[0034] A previous timestamp T-1 is obtained, and the target historical time series data set is searched according to the previous timestamp T-1 to obtain the historical recommended strategy corresponding to the previous timestamp T-1.

[0035] The historical probability of the historical state corresponding to the historical recommended strategy is obtained, wherein the historical probability of the historical state comprises a historical accepted probability or a historical unaccepted probability.

[0036] Optionally, the determining of the recommended strategy as a target recommended strategy according to the comparison result and the recommending of the target recommended strategy comprise:

[0037] If the comparison result is that the accepted probability is greater than the accepted probability threshold, the recommended strategy is determined as the target recommended strategy, and the target recommended strategy is recommended.

[0038] If the comparison result is that the accepted probability is less than the accepted probability threshold, the recommended strategy is not recommended.

[0039] Embodiments of the present application also provide a strategy recommendation device, comprising:

[0040] The training module is configured to obtain a training data set and train an initial prediction model based on the training data set to obtain a target prediction model.

[0041] The input module is configured to obtain a to-be-recommended strategy and input the to-be-recommended strategy into the target prediction model.

[0042] The output module is configured to output an acceptance probability of the to-be-recommended strategy based on a historical probability of a historical state of a previous historical recommended strategy of the to-be-recommended strategy by the target prediction model.

[0043] The comparison module is configured to compare the acceptance probability with the acceptance probability threshold to obtain a comparison result.

[0044] The recommendation module is configured to determine the to-be-recommended strategy as a target recommended strategy according to the comparison result and recommend the target recommended strategy.

[0045] Embodiments of the present application also provide an electronic device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the method described above when executing the computer program.

[0046] Embodiments of the present application also provide a computer-readable storage medium including a stored computer program, wherein the computer-readable storage medium controls a device where the computer-readable storage medium is located to execute the method described above when the computer program runs.

[0047] Embodiments of the present application have the following technical effects:

[0048] The above technical solutions of the present application predict an acceptance probability of a to-be-recommended strategy based on a historical state of a previous historical recommended strategy and a historical probability corresponding to the historical state, and when the predicted acceptance probability of the to-be-recommended strategy is small, the to-be-recommended strategy is not recommended to a user, thereby reducing the influence of strategy recommendation on the user and improving the satisfaction of the user.

[0049] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a flowchart of a strategy recommendation method provided by an embodiment of the present application;

[0051] Figure 2 is a structural schematic diagram of a strategy recommendation device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0052] Embodiments of the present application are described below in detail with reference to the accompanying drawings, examples of which are shown in the drawings, wherein the same or similar notations represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by reference to the drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0053] When a user obtains some recommended strategies sent by a service end (cloud, automatic driving system, etc.) based on a terminal (mobile phone, etc.), if the user rejects the recommended strategies, if the user is performing other operations when the recommended strategies are recommended to the user, the recommended strategies recommended to the user will interfere with the normal work of the user. In order to solve this problem, embodiments of the present application provide the following technical solutions:

[0054] As shown in Figure 1 The embodiments of the present application provide a strategy recommendation method, comprising:

[0055] Step S11: obtaining a training data set, and training an initial prediction model based on the training data set to obtain a target prediction model;

[0056] Specifically, the obtaining of the training data set and the training of the initial prediction model based on the training data set to obtain the target prediction model comprises:

[0057] Based on a preset rule, the training data set is classified to obtain a plurality of historical data sets; each of the historical data sets comprises a plurality of historical data; each of the historical data comprises a historical recommended strategy, a historical state corresponding to the historical recommended strategy, a historical probability of the historical state, and a timestamp;

[0058] The timestamp of each of the historical data in each of the historical data sets is obtained;

[0059] Based on the timestamp, the plurality of historical data in each of the historical data sets is sorted to obtain a plurality of historical time series data sets;

[0060] The initial prediction model is trained based on the plurality of historical time series data sets to obtain the target prediction model.

[0061] The embodiments of the present application construct an initial prediction model, and then obtain reaction result data of users with similar behaviors (the same personal preferences, etc.) to recommendations (accepting recommendations or not accepting recommendations) in a period of time (the length of the period of time can be set according to actual needs). According to the reaction result data, a training data set is obtained.

[0062] Specifically, the preset rule related to the embodiments of the present application can be a scene or a device, and can be set or adjusted according to actual needs of a user.

[0063] For example, 1) the preset rule is a scene, and the training data is classified based on the scene, for example, the training data is classified based on weather, road conditions, parking, refueling, etc.

[0064] 2) the preset rule can also be a device, and the training data is classified based on the device, for example, the training data is classified based on a steering wheel, an accelerator, and a door, etc.

[0065] In the embodiments of the present application, each historical data can include a historical recommendation strategy, a historical state corresponding to the historical recommendation strategy, a historical probability corresponding to the historical state, and a timestamp.

[0066] For example, a certain historical recommendation strategy is refueling, a historical state of refueling, a historical probability corresponding to the historical state, and a time when the historical recommendation strategy occurs.

[0067] According to the timestamps of each historical data, each historical data is sorted in chronological order, and then it can be determined whether the historical recommendation strategy needs to be recommended to the user based on whether the previous historical recommendation strategy is accepted or not accepted.

[0068] In an optional embodiment of the present application, the training of the initial prediction model based on a plurality of historical time series data sets to obtain a target prediction model includes:

[0069] Obtaining a historical recommendation strategy;

[0070] According to the historical recommendation strategy, a target historical time series data set is determined;

[0071] Based on a plurality of historical states in the target historical time series data set, a target historical state chain is obtained; wherein the historical state includes a historical acceptance state or a historical non-acceptance state;

[0072] Based on the target historical state chain, the acceptance probability of the historical recommendation strategy is calculated.

[0073] In the embodiments of the present application, in the process of training the initial prediction model, a historical recommendation strategy is constantly obtained, for example, historical recommendation of refueling or historical recommendation of parking space, etc.

[0074] When the historical to-be-recommended strategy is refueling, a category to which the refueling belongs is determined, for example, a refueling scenario, then a plurality of historical time-series data sets are searched according to the refueling scenario, a historical time-series data set corresponding to the refueling scenario is searched, and the historical time-series data set is determined as a target historical time-series data set.

[0075] Then, a target historical state chain is obtained based on a plurality of historical states in the target historical time-series data set; wherein the target historical state chain can be represented based on a Markov chain.

[0076] For example, the historical time-series data set corresponding to the refueling scenario includes 10 recommended historical recommended strategies, after sorting according to timestamps, the following can be obtained:

[0077] The first historical recommended strategy is refueling, the second historical recommended strategy is refueling, and the tenth historical recommended strategy is refueling.

[0078] The historical states corresponding to the 10 recommended historical recommended strategies are historical acceptance state (user accepts refueling), historical acceptance state, historical non-acceptance state (user does not accept refueling), historical non-acceptance state, historical acceptance state, historical non-acceptance state, historical non-acceptance state, historical acceptance state, historical non-acceptance state, and historical acceptance state.

[0079] Therefore, the target historical state chain can be obtained based on the historical states corresponding to the 10 recommended historical recommended strategies in turn:

[0080] Historical acceptance state→historical acceptance state→historical non-acceptance state→historical non-acceptance state→historical acceptance state→historical non-acceptance state→historical non-acceptance state→historical acceptance state→historical non-acceptance state→historical acceptance state.

[0081] Among them, the embodiment of the application shows that the state transition process of the first historical state is from the historical acceptance state to the historical acceptance state.

[0082] The state transition process of the second historical state is from the historical acceptance state to the historical non-acceptance state.

[0083] By analogy, the target historical state chain shows the state transition processes of the first nine historical states.

[0084] And the tenth historical state is used to predict the eleventh historical state and whether it will be accepted by the user, and is represented by an acceptance probability; when the eleventh historical state is determined, the state transition process of the tenth historical state can also be determined, and by analogy, the target historical time-series data set and the target historical state chain of each historical to-be-recommended strategy can be obtained.

[0085] In an optional embodiment of this application, calculating the acceptance probability of the historical recommendation strategy based on the target historical state chain includes:

[0086] A first value is calculated based on the target historical state chain; wherein the first value is used to characterize the number of historical acceptance states included in the target historical state chain;

[0087] A second value is calculated based on the target historical state chain; wherein the second value is used to characterize the number of the historical unacceptable states included in the target historical state chain;

[0088] Based on the target historical state chain, the first value, and the second value, the target matrix of the target historical state chain is calculated.

[0089] The acceptance probability of the historical recommendation strategy is calculated based on the target matrix of the target historical state chain and the historical probability of the historical state of the previous historical recommendation strategy.

[0090] For example: 1) When the target historical state chain is:

[0091] Acceptance state → Acceptance state → Disapproval state → Disapproval state → Acceptance state → Disapproval state → Disapproval state → Acceptance state → Disapproval state → Acceptance state;

[0092] Therefore, the number of historical acceptance states is 5; that is, the first value is 5.

[0093] The number of states that history does not accept is 5; that is, the second value is 5.

[0094] 2) When the target historical state chain is:

[0095] Acceptance state → Disapproval state → Disapproval state → Disapproval state → Acceptance state → Disapproval state → Disapproval state → Disapproval state → Acceptance state;

[0096] Therefore, the number of historical acceptance states is 3; that is, the first value is 3.

[0097] The number of states that history does not accept is 7; that is, the second value is 7.

[0098] By analogy, the number of historical accepting states and the number of historical rejecting states in any target historical state chain can be calculated.

[0099] Then, the target matrix is ​​calculated based on the first and second values, and the acceptance probability of the historical recommendation strategy is calculated based on the target matrix and the historical probability of the historical state of the previous historical recommendation strategy.

[0100] In an optional embodiment of this application, calculating the target matrix of the target historical state chain based on the target historical state chain, the first value, and the second value includes:

[0101] A first transition value is calculated; wherein the first transition value is used to characterize the number of transitions from the historical accepting state to the historical accepting state in the target historical state chain;

[0102] A second transition value is calculated; wherein the second transition value is used to characterize the number of states in the target historical state chain that transition from the historical accepting state to the historical non-accepting state;

[0103] A third transition value is calculated; wherein, the third transition value is used to characterize the number of states in the target historical state chain that transition from the historical unacceptable state to the historical accepting state;

[0104] A fourth transition value is calculated; wherein, the fourth transition value is used to characterize the number of transitions from the historical unacceptable state to the historical unacceptable state in the target historical state chain;

[0105] Based on the first transition value, the second transition value, the third transition value, and the fourth transition value, the target matrix of the target historical state chain is calculated.

[0106] For example: when the target historical state chain is:

[0107] Accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Not accepting state → Not accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state;

[0108] Therefore, the number of historical acceptance states is 12; that is, the first value is 12.

[0109] The number of states that history does not accept is 8; that is, the second value is 8.

[0110] Of the 12 historical acceptance states, 6 transitioned from one historical acceptance state to another, meaning the first transition value is 6.

[0111] Therefore, the first parameter = 6 / 12 = 0.5.

[0112] Of the 12 historical acceptance states, 6 transitioned from historical acceptance states to historical non-acceptance states, which means the second transition value is 6.

[0113] Therefore, the second parameter = 6 / 12 = 0.5.

[0114] Of the 8 historical rejection states, 6 transitioned from historical rejection states to historical acceptance states, which means the third transition value is 6.

[0115] Therefore, the third parameter = 6 / 8 = 0.75.

[0116] Of the 8 historical rejection states, 2 transitioned from one historical rejection state to another, meaning the fourth transition value is 2.

[0117] Therefore, the fourth parameter = 2 / 8 = 0.25.

[0118] Target matrix = ;

[0119] Then the target matrix = .

[0120] In an optional embodiment of this application, obtaining the historical probability of the historical state of the historical recommendation strategy includes:

[0121] Obtain the timestamp T of the historical recommendation strategy;

[0122] Obtain the previous timestamp T-1, and search the target historical time series dataset based on the previous timestamp T-1 to obtain the historical recommendation strategy corresponding to the previous timestamp T-1;

[0123] Obtain the historical probability of the historical state corresponding to the historical recommendation strategy; wherein the historical probability of the historical state includes the historical acceptance probability or the historical non-acceptance probability.

[0124] For example: when the target historical state chain is:

[0125] Accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Not accepting state → Not accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state → Not accepting state → Accepting state → Accepting state;

[0126] The historical recommendation strategy to be recommended is the twenty-first historical recommendation strategy. Assuming the timestamp of the twenty-first historical recommendation strategy is T, the timestamp of the twenty-second historical recommendation strategy is T-1, the timestamp of the nineteenth historical recommendation strategy is T-2, and so on, the timestamp of the first historical recommendation strategy is T-20.

[0127] Similarly, based on the timestamps of the historical recommendation strategies, the timestamps of the previous historical recommendation strategies can be determined, thereby obtaining the historical probability of the historical state of the previous historical recommendation strategies.

[0128] In the embodiments of this application, the previous adjacent timestamp is determined based on the timestamp of the strategy to be recommended, and the previous adjacent historical recommendation strategy is determined based on the previous adjacent timestamp. Then, the acceptance probability of the strategy to be recommended is predicted based on the historical state and historical probability of the previous adjacent historical recommendation strategy. The data is easy to obtain, the calculation is simple, and the computational cost is reduced.

[0129] For example, if the historical acceptance probability of a previous recommended refueling station is 0.8, then the historical rejection probability is 0.2.

[0130] Then it should be recommended to cheer. = *Target matrix;

[0131] For example: target matrix = ;

[0132] Then it should be recommended to cheer. = *

[0133] Therefore, the recommended refueling station = ;

[0134] That is, the probability of accepting the recommendation for refueling is 0.55.

[0135] Step S12: Obtain the strategy to be recommended and input the strategy to be recommended into the target prediction model;

[0136] In the embodiments of this application, assuming that the target prediction model is applied to autonomous driving, during the operation of the vehicle, the autonomous driving system continuously generates a strategy to be recommended based on the vehicle's operating status data, and sends the strategy to be recommended to the target prediction model.

[0137] The target prediction model first determines the target's historical time series dataset based on the policy to be predicted, and then obtains the target's historical state chain based on the target's historical time series dataset.

[0138] The target matrix is ​​calculated based on the target's historical state chain.

[0139] Step S13: The target prediction model outputs the acceptance probability of the strategy to be recommended based on the historical probability of the historical state of the previous historical recommendation strategy.

[0140] In embodiments of this application, the target prediction model outputs the acceptance probability of the strategy to be recommended based on the target matrix and the historical probability of the historical state of the previous historical recommendation strategy.

[0141] Step S14: Compare the acceptance probability with the acceptance probability threshold to obtain the comparison result;

[0142] It should be noted that the acceptance probability threshold can be set according to the user's actual needs, and the embodiments of this application do not impose specific limitations on it.

[0143] Step S15: Based on the comparison results, determine the strategy to be recommended as the target recommendation strategy, and recommend the target recommendation strategy.

[0144] Specifically, the step of determining the strategy to be recommended as the target recommendation strategy based on the comparison result, and recommending the target recommendation strategy, includes:

[0145] If the comparison result shows that the acceptance probability is greater than the acceptance probability threshold, then the strategy to be recommended is determined as the target recommendation strategy, and the target recommendation strategy is recommended.

[0146] If the comparison result shows that the acceptance probability is less than the acceptance probability threshold, then the strategy to be recommended will not be recommended.

[0147] In the embodiments of this application, by setting or adjusting the acceptance probability threshold, different numbers of recommendation strategies can be selected for different acceptance probability thresholds, thereby meeting the needs of different users and improving user satisfaction.

[0148] In an optional embodiment of this application, assuming the acceptance probability threshold is 0.8, when the acceptance probability of the strategy to be recommended obtained by the above calculation is 0.55, the acceptance probability of the strategy to be recommended is less than the acceptance probability threshold. Therefore, it is predicted that the strategy to be recommended will not be accepted by the user, so the strategy to be recommended will not be recommended.

[0149] Then, the target prediction model stores the relevant data of the strategy to be recommended as historical data, which is used to update the data in the target historical time series dataset corresponding to the strategy to be recommended, and can also predict whether the next strategy to be recommended will be accepted.

[0150] Specifically, the relevant data corresponding to the strategy to be recommended includes:

[0151] Timestamp;

[0152] Historical status: Not accepting;

[0153] The probability of not accepting is 0.45, and the probability of accepting is 0.55.

[0154] Suppose that the target historical state chain corresponding to the strategy to be recommended is: accept state → accept state → accept state → do not accept state → accept state → do not accept state → do not accept state → accept state → do not accept state → accept state → accept state → do not accept state → accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state → accept state → accept state;

[0155] Then, the target historical state chain corresponding to the next recommended strategy is: accept state → accept state → accept state → do not accept state → accept state → do not accept state → do not accept state → accept state → do not accept state → accept state → accept state → do not accept state → accept state → accept state → do not accept state → accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state → accept state → do not accept state.

[0156] In the embodiments of this application, based on the historical state of the previous historical recommendation strategy and the historical probability corresponding to the historical state, the acceptance probability of the recommendation strategy to be accepted by the user is predicted. When the predicted acceptance probability of the recommendation strategy to be recommended is small, the recommendation strategy to be recommended is not recommended to the user, thereby reducing the impact of strategy recommendation on the user and improving user satisfaction.

[0157] The embodiments of this application can be implemented based on the following methods:

[0158] For example, 1) If the autonomous driving system detects that the vehicle speed is gradually decreasing, it predicts that the vehicle may need to stop. Therefore, the autonomous driving system obtains nearby parking spaces and recommends the most suitable parking space to the user.

[0159] 2) The target prediction model obtains the recommended parking space strategy.

[0160] 3) The target prediction model finds the target historical time series dataset that matches the parking space to be recommended.

[0161] 4) The target prediction model obtains the target's historical state chain based on the target's historical time series dataset;

[0162] For example: The target historical state chain is:

[0163] Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Accepting state → Non-accepting state → Accepting state → Non-accepting state → Non-accepting state → Accepting state → Accepting state → Accepting state;

[0164] 5) The target prediction model calculates the target matrix based on the target's historical state chain.

[0165] The number of historical acceptance states is 17; that is, the first value is 17.

[0166] The number of states that history does not accept is 3; that is, the second value is 3.

[0167] Of the 17 historical acceptance states, 14 transitioned from one historical acceptance state to another, meaning the first transition value is 14.

[0168] Therefore, the first parameter = 14 / 17 = 0.82.

[0169] Of the 17 historical acceptance states, 3 transitioned from historical acceptance states to historical non-acceptance states, which means the second transition value is 3.

[0170] Therefore, the second parameter = 3 / 17 = 0.18.

[0171] Of the three historical rejection states, two transitioned from historical rejection states to historical acceptance states, meaning the fourth transition value is 1.

[0172] Therefore, the third parameter = 2 / 3 = 0.67.

[0173] Of the three historical rejection states, the number of transitions from one historical rejection state to another is 1, which means the fourth transition value is 1.

[0174] Therefore, the fourth parameter = 1 / 3 = 0.67.

[0175] Then the target matrix = .

[0176] If the historical acceptance probability of a previous recommended refueling service is 0.8, then the historical rejection probability is 0.2.

[0177] Then the parking space to be recommended = * ;

[0178] Then the parking space to be recommended = ;

[0179] That is, the acceptance probability of the recommended parking space is 0.79.

[0180] Assuming the acceptance probability threshold is 0.78, then the acceptance probability 0.79 is greater than the acceptance probability threshold of 0.78. Therefore, the parking space to be recommended is determined as the target recommended parking space, and the parking space is recommended to the user. The user can receive the recommended parking space through a terminal set on their car or a smartphone or other device.

[0181] If the user accepts the recommended parking space, the historical status of the parking space to be recommended will be recorded as: accepted status.

[0182] like Figure 2 As shown, embodiments of this application also provide a strategy recommendation device 20, comprising:

[0183] Training module 21 is used to acquire a training dataset and train an initial prediction model based on the training dataset to obtain a target prediction model;

[0184] Input module 22 is used to obtain the strategy to be recommended and input the strategy to be recommended into the target prediction model;

[0185] Output module 23 is used to output the acceptance probability of the recommended strategy based on the historical probability of the historical state of the previous historical recommended strategy of the recommended strategy.

[0186] Comparison module 24 is used to compare the acceptance probability with the acceptance probability threshold to obtain a comparison result;

[0187] The recommendation module 25 is used to determine the strategy to be recommended as the target recommendation strategy based on the comparison result, and to recommend the target recommendation strategy.

[0188] Optionally, obtaining a training dataset and training an initial prediction model based on the training dataset to obtain a target prediction model includes:

[0189] Based on preset rules, the training dataset is classified to obtain multiple historical datasets; wherein each historical dataset includes multiple historical data; each historical data includes a historical recommendation strategy, a historical state corresponding to the historical recommendation strategy, a historical probability of the historical state, and a timestamp;

[0190] Obtain the timestamp of each historical data in each historical dataset;

[0191] Based on the timestamp, sort the multiple historical data in each historical dataset to obtain multiple historical time series datasets;

[0192] The initial prediction model is trained based on multiple historical time-series datasets to obtain the target prediction model.

[0193] Optionally, training the initial prediction model based on multiple historical time-series datasets to obtain the target prediction model includes:

[0194] Obtain historical recommendation strategies;

[0195] Based on the aforementioned historical recommendation strategy, the target historical time series dataset is determined;

[0196] Based on multiple historical states in the target historical time series dataset, a target historical state chain is obtained; wherein, the historical states include historical accepting states or historical non-accepting states;

[0197] Based on the target historical state chain, the acceptance probability of the historical recommended strategy is calculated.

[0198] Optionally, calculating the acceptance probability of the historical recommendation strategy based on the target historical state chain includes:

[0199] A first value is calculated based on the target historical state chain; wherein the first value is used to characterize the number of historical acceptance states included in the target historical state chain;

[0200] A second value is calculated based on the target historical state chain; wherein the second value is used to characterize the number of the historical unacceptable states included in the target historical state chain;

[0201] Based on the target historical state chain, the first value, and the second value, the target matrix of the target historical state chain is calculated.

[0202] The acceptance probability of the historical recommendation strategy is calculated based on the target matrix of the target historical state chain and the historical probability of the historical state of the previous historical recommendation strategy.

[0203] Optionally, calculating the target matrix of the target historical state chain based on the target historical state chain, the first value, and the second value includes:

[0204] A first transition value is calculated; wherein the first transition value is used to characterize the number of transitions from the historical accepting state to the historical accepting state in the target historical state chain;

[0205] A second transition value is calculated; wherein the second transition value is used to characterize the number of states in the target historical state chain that transition from the historical accepting state to the historical non-accepting state;

[0206] A third transition value is calculated; wherein, the third transition value is used to characterize the number of states in the target historical state chain that transition from the historical unacceptable state to the historical accepting state;

[0207] A fourth transition value is calculated; wherein, the fourth transition value is used to characterize the number of transitions from the historical unacceptable state to the historical unacceptable state in the target historical state chain;

[0208] Based on the first transition value, the second transition value, the third transition value, and the fourth transition value, the target matrix of the target historical state chain is calculated.

[0209] Optionally, obtaining the historical probability of the historical state of the historical recommendation strategy includes:

[0210] Obtain the timestamp T of the historical recommendation strategy;

[0211] Obtain the previous timestamp T-1, and search the target historical time series dataset based on the previous timestamp T-1 to obtain the historical recommendation strategy corresponding to the previous timestamp T-1;

[0212] Obtain the historical probability of the historical state corresponding to the historical recommendation strategy; wherein the historical probability of the historical state includes the historical acceptance probability or the historical non-acceptance probability.

[0213] Optionally, determining the strategy to be recommended as the target recommendation strategy based on the comparison result, and recommending the target recommendation strategy, includes:

[0214] If the comparison result shows that the acceptance probability is greater than the acceptance probability threshold, then the strategy to be recommended is determined as the target recommendation strategy, and the target recommendation strategy is recommended.

[0215] If the comparison result shows that the acceptance probability is less than the acceptance probability threshold, then the strategy to be recommended will not be recommended.

[0216] Embodiments of this application also provide an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the method described above.

[0217] Embodiments of this application also provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the method described above.

[0218] Furthermore, other configurations and functions of the apparatus in the embodiments of this application are known to those skilled in the art, and will not be described in detail here to reduce redundancy.

[0219] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0220] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0221] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0222] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0223] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0224] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0225] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0226] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A strategy recommendation method, characterized in that, include: Obtain a training dataset and train the initial prediction model based on the training dataset to obtain the target prediction model; Obtain the strategy to be recommended and input the strategy to be recommended into the target prediction model; The target prediction model outputs the acceptance probability of the recommended strategy based on the historical probability of the historical state of the previous historical recommendation strategy. The acceptance probability is compared with the acceptance probability threshold to obtain the comparison result; Based on the comparison results, the strategy to be recommended is determined as the target recommendation strategy, and the target recommendation strategy is recommended. The step of obtaining a training dataset and training an initial prediction model based on the training dataset to obtain a target prediction model includes: Based on preset rules, the training dataset is classified to obtain multiple historical datasets; wherein each historical dataset includes multiple historical data; each historical data includes a historical recommendation strategy, a historical state corresponding to the historical recommendation strategy, a historical probability of the historical state, and a timestamp; Obtain the timestamp of each historical data in each historical dataset; Based on the timestamp, sort the multiple historical data in each historical dataset to obtain multiple historical time series datasets; The initial prediction model is trained based on multiple historical time-series datasets to obtain the target prediction model; the process of training the initial prediction model based on multiple historical time-series datasets to obtain the target prediction model includes: Obtain historical recommendation strategies; Based on the aforementioned historical recommendation strategy, the target historical time series dataset is determined; Based on multiple historical states in the target historical time series dataset, a target historical state chain is obtained; wherein, the historical states include historical accepting states or historical non-accepting states; Based on the target historical state chain, the acceptance probability of the historical recommended strategy is calculated.

2. The method according to claim 1, characterized in that, The step of calculating the acceptance probability of the historical recommendation strategy based on the target historical state chain includes: A first value is calculated based on the target historical state chain; wherein the first value is used to characterize the number of historical acceptance states included in the target historical state chain; A second value is calculated based on the target historical state chain; wherein the second value is used to characterize the number of the historical unacceptable states included in the target historical state chain; Based on the target historical state chain, the first value, and the second value, the target matrix of the target historical state chain is calculated. The acceptance probability of the historical recommendation strategy is calculated based on the target matrix of the target historical state chain and the historical probability of the historical state of the previous historical recommendation strategy.

3. The method according to claim 2, characterized in that, The step of calculating the target matrix of the target historical state chain based on the target historical state chain, the first value, and the second value includes: A first transition value is calculated; wherein the first transition value is used to characterize the number of transitions from the historical accepting state to the historical accepting state in the target historical state chain; A second transition value is calculated; wherein the second transition value is used to characterize the number of states in the target historical state chain that transition from the historical accepting state to the historical non-accepting state; A third transition value is calculated; wherein, the third transition value is used to characterize the number of states in the target historical state chain that transition from the historical unacceptable state to the historical accepting state; A fourth transition value is calculated; wherein, the fourth transition value is used to characterize the number of transitions from the historical unacceptable state to the historical unacceptable state in the target historical state chain; Based on the first transition value, the second transition value, the third transition value, and the fourth transition value, the target matrix of the target historical state chain is calculated.

4. The method according to claim 1, characterized in that, The historical probability of obtaining the historical state of the historical recommendation strategy includes: Obtain the timestamp T of the historical recommendation strategy; Obtain the previous timestamp T-1, and search the target historical time series dataset based on the previous timestamp T-1 to obtain the historical recommendation strategy corresponding to the previous timestamp T-1; Obtain the historical probability of the historical state corresponding to the historical recommendation strategy; wherein the historical probability of the historical state includes the historical acceptance probability or the historical non-acceptance probability.

5. The method according to claim 1, characterized in that, The step of determining the strategy to be recommended as the target recommendation strategy based on the comparison result, and recommending the target recommendation strategy, includes: If the comparison result shows that the acceptance probability is greater than the acceptance probability threshold, then the strategy to be recommended is determined as the target recommendation strategy, and the target recommendation strategy is recommended. If the comparison result shows that the acceptance probability is less than the acceptance probability threshold, then the strategy to be recommended will not be recommended.

6. A strategy recommendation device, characterized in that, include: The training module is used to acquire a training dataset and train an initial prediction model based on the training dataset to obtain a target prediction model. The input module is used to obtain the strategy to be recommended and input the strategy to be recommended into the target prediction model; The output module is used to output the acceptance probability of the recommended strategy based on the historical probability of the historical state of the previous historical recommended strategy of the recommended strategy. The comparison module is used to compare the acceptance probability with the acceptance probability threshold to obtain a comparison result; The recommendation module is used to determine the strategy to be recommended as the target recommendation strategy based on the comparison results, and to recommend the target recommendation strategy. The step of obtaining a training dataset and training an initial prediction model based on the training dataset to obtain a target prediction model includes: Based on preset rules, the training dataset is classified to obtain multiple historical datasets; wherein each historical dataset includes multiple historical data; each historical data includes a historical recommendation strategy, a historical state corresponding to the historical recommendation strategy, a historical probability of the historical state, and a timestamp; Obtain the timestamp of each historical data in each historical dataset; Based on the timestamp, sort the multiple historical data in each historical dataset to obtain multiple historical time series datasets; The initial prediction model is trained based on multiple historical time-series datasets to obtain the target prediction model; the process of training the initial prediction model based on multiple historical time-series datasets to obtain the target prediction model includes: Obtain historical recommendation strategies; Based on the aforementioned historical recommendation strategy, the target historical time series dataset is determined; Based on multiple historical states in the target historical time series dataset, a target historical state chain is obtained; wherein, the historical states include historical accepting states or historical non-accepting states; Based on the target historical state chain, the acceptance probability of the historical recommended strategy is calculated.

7. An electronic device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the method as claimed in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the method as described in any one of claims 1 to 5.

Images