Target-subject determination method and apparatus, and device and storage medium

Abstract

A target-subject determination method and apparatus, and a device and a storage medium. The determination method comprises: on the basis of a plurality of performance characteristics of each participant subject in the current competition, respectively determining the degrees of contribution of the performance characteristics to a competition prediction result outputted from a competition result prediction model when model input data of the competition result prediction model includes the performance characteristics (S101); summarizing the degrees of contribution respectively corresponding to the plurality of performance characteristics, so as to obtain a degree-of-contribution assessment result of each participant subject in the current competition (S102); and on the basis of the degree-of-contribution assessment result, determining a target subject of the current competition from among the participant subjects (S103). Therefore, a more fair, objective and rational basis is provided for the assessment of a target subject, such that a target-subject assessment result has the characteristics of simplicity, transparency and high interpretability, which are conducive to helping organizers of various competitions assess the competition performance of each participant subject more comprehensively and objectively.

Description

Method, device and equipment for determining target object and storage medium

[0001] Cross-reference to Related Applications

[0002] This application claims priority to Chinese Patent Application No. 202410979559.7, filed on July 19, 2024, entitled “Method, device and equipment for determining target object and storage medium”, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0003] The present disclosure relates to the technical field of data processing, in particular, to a method, device and equipment for determining a target object and a storage medium. BACKGROUND

[0004] After the end of a competition, it is often necessary to select a target object with the best performance in the competition from all participating objects according to the performance of each participating object in the competition. For example, the target object can be the MVP (Most Valuable Player) player in the competition.

[0005] The existing target object selection scheme can be divided into two categories: the first category is to obtain the performance score of each participating object in the competition by weighted summation according to the performance characteristics of the participating object in the competition and the weight coefficient corresponding to each performance characteristic, and to select the participating object with the highest performance score as the target object with the best performance in the competition (mainly used in online competitions); the second category is to determine the participating object with the highest number of expert votes as the target object with the best performance in the competition by expert voting (mainly used in offline competitions).

[0006] However, the above two categories of target object selection schemes essentially rely on human experience in terms of the weight coefficient corresponding to different performance characteristics and the specific voting of experts, so that the final target object selection result cannot be explained by more objective data, resulting in the defects that the target object selection result is greatly interfered by human factors and has poor interpretability. SUMMARY

[0007] Therefore, the present disclosure provides a method, device and equipment for determining a target object and a storage medium, which improves the existing target object selection method and provides a more fair, objective and reasonable basis for the selection of the target object, so that the target object selection result has the characteristics of simplicity, transparency and strong interpretability, which is beneficial to help various event organizers more comprehensively and objectively evaluate the performance of each participating object in the competition.

[0008] In order to make the above objectives, characteristics and advantages of the present disclosure more obvious and easily understood, the following preferred embodiments are specifically described below with reference to the attached drawings.

[0009] In a first aspect, the embodiments of the present disclosure provide a target object determination method, which comprises the following steps:

[0010] According to a plurality of performance characteristics of each contestant in the current competition, the contribution degree of each performance characteristic to the competition prediction result output by the competition result prediction model is determined under the condition that the performance characteristic is contained in the model input data of the competition result prediction model.

[0011] The contribution degrees corresponding to the plurality of performance characteristics are summarized to obtain a contribution degree evaluation result of each contestant in the current competition.

[0012] According to the contribution degree evaluation result, a target object of the current competition is determined from the contestants.

[0013] In a second aspect, the embodiments of the present disclosure provide a target object determination device, which comprises:

[0014] The contribution allocation module is configured to determine, according to a plurality of performance characteristics of each contestant in the current competition, the contribution degree of each performance characteristic to the competition prediction result output by the competition result prediction model under the condition that the performance characteristic is contained in the model input data of the competition result prediction model.

[0015] The statistics module is configured to summarize the contribution degrees corresponding to the plurality of performance characteristics to obtain a contribution degree evaluation result of each contestant in the current competition.

[0016] The selection module is configured to determine a target object of the current competition from the contestants according to the contribution degree evaluation result.

[0017] In a third aspect, the embodiments of the present disclosure provide a computer device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the target object determination method when executing the computer program.

[0018] In a fourth aspect, the embodiments of the present disclosure provide a computer readable storage medium, which stores a computer program, and the computer program is executable on a processor to execute the steps of the target object determination method.

[0019] The technical solutions provided by the embodiments of the present disclosure can include the following beneficial effects:

[0020] The method for determining a target object provided in the embodiments of the present disclosure includes: determining, according to a plurality of performance characteristics of each contestant object in the current competition, a contribution degree of each performance characteristic to a competition prediction result output by a competition result prediction model under a condition that the performance characteristic is included in model input data of the competition result prediction model; collecting the contribution degrees corresponding to the plurality of performance characteristics respectively to obtain a contribution degree evaluation result of each contestant object in the current competition; and determining the target object of the current competition from the contestant objects according to the contribution degree evaluation result. In this way, the present disclosure provides a more fair, objective and reasonable basis for selection of the target object, so that the selection result of the target object has the characteristics of simplicity, transparency and strong interpretability, and is conducive to helping various event organizers to more comprehensively and objectively evaluate the performance of each contestant object in the competition. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the present disclosure, and therefore should not be considered as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained from these drawings without creative labor.

[0022] FIG. 1 shows a flowchart of a method for determining a target object according to an embodiment of the present disclosure;

[0023] FIG. 2 shows a flowchart of a method for training a competition result prediction model according to an embodiment of the present disclosure;

[0024] FIG. 3 shows a flowchart of a method for optimizing a plurality of competition characteristics needed to be referred to when selecting a target object according to an embodiment of the present disclosure;

[0025] FIG. 4 shows a structural diagram of a device for determining a target object according to an embodiment of the present disclosure;

[0026] FIG. 5 is a structural diagram of an electronic device 500 according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] In order to make the purposes, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings in the embodiments of the present disclosure. It should be understood that the drawings in the present disclosure only serve the purpose of illustration and description, and are not used to limit the protection scope of the present disclosure. In addition, it should be understood that the schematic drawings are not drawn according to the actual proportions. The flowcharts used in the present disclosure show the operations implemented according to some embodiments of the present disclosure. It should be understood that the operations of the flowcharts can not be implemented in sequence, and the steps without logical context relationship can be reversed in sequence or implemented simultaneously. In addition, one or more other operations can be added to the flowcharts or one or more operations can be removed from the flowcharts under the guidance of the present disclosure.

[0028] In addition, the described embodiments are only part of the embodiments of the present disclosure, not all the embodiments. The components of the embodiments of the present disclosure described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present disclosure provided in the drawings is not intended to limit the scope of the claimed present disclosure, but only represents selected embodiments of the present disclosure. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present disclosure.

[0029] It should be noted that the term "comprising" will be used in the embodiments of the present disclosure to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0030] The existing two types of target object selection schemes are essentially dependent on human experience in terms of the weight coefficients corresponding to different performance characteristics and the specific voting of experts, so that the final target object selection result cannot be explained by more objective data, and the target object selection result has the defects of being greatly interfered by human factors and poor interpretability.

[0031] Based on this, the embodiments of the present disclosure provide a target object determination method and device, equipment and storage medium, by improving the existing target object selection method, a more fair, objective and reasonable basis is provided for the selection of target objects, so that the target object selection result has the characteristics of simple transparency and strong interpretability, which is beneficial to help various event organizers more comprehensively and objectively evaluate the performance of each participating object in the competition.

[0032] In one embodiment of this disclosure, a method for determining a target object can run on a terminal device or a server. The terminal device can be a local terminal device. When the method for determining the target object runs on a server, it can be implemented and executed based on a cloud interaction system, which includes a server and client devices (i.e., terminal devices).

[0033] To facilitate understanding of the embodiments of this disclosure, a method, apparatus, device, and storage medium for determining a target object provided by the embodiments of this disclosure will be described in detail below.

[0034] Referring to Figure 1, which illustrates a flowchart of a method for determining a target object according to an embodiment of this disclosure, the method includes steps S101-S103; specifically:

[0035] S101, based on multiple performance characteristics of each participant in this match, determine the degree of contribution of the performance characteristics to the match prediction results output by the match prediction model, provided that the model input data of the match result prediction model includes the performance characteristics.

[0036] S102, the contribution levels corresponding to the multiple performance characteristics are summarized to obtain the contribution evaluation result of each participant in this game.

[0037] S103, Based on the contribution evaluation results, determine the target of this match from the participants.

[0038] The method for determining the target object provided in this disclosure determines the contribution of each performance feature to the prediction result output by the prediction model, given that the model input data includes the performance features, based on multiple performance features of each participant in the match. The contribution levels corresponding to each performance feature are then summarized to obtain a contribution evaluation result for each participant in the match. Based on the contribution evaluation result, the target object for the match is determined from among the participants. This disclosure provides a fairer, more objective, and reasonable basis for the selection of target objects, making the selection results simple, transparent, and highly interpretable. This helps various event organizers to more comprehensively and objectively evaluate the performance of each participant.

[0039] The following is an exemplary description of each step in the method for determining the target object provided in the embodiments of this disclosure:

[0040] S101, respectively determine, according to a plurality of performance features of each contestant in the current competition, a contribution degree of the performance features to a competition prediction result output by the competition result prediction model under a condition that the performance features are contained in model input data of the competition result prediction model.

[0041] Here, the current competition represents a competition that needs to select a target object; wherein the current competition can belong to an online competition (such as a game match in a game), or can belong to an offline competition (such as a basketball match played offline), and the target object can represent a contestant with the best performance in the current competition (such as an MVP of the current competition), or can represent a contestant who makes the highest contribution to his / her team in the current competition. The specific competition type to which the current competition belongs and the specific meaning represented by the target object in the current competition are not limited in the embodiments of the present disclosure.

[0042] Here, after the current competition ends, for each contestant, game performance data of the contestant in the current competition can be obtained from competition data recorded in the current competition, and then, according to a plurality of competition features that need to be referred to when determining the target object, target game performance data matched with each competition feature can be extracted from the game performance data of the contestant as a plurality of performance features of the contestant in the current competition; that is, one performance feature can correspond to one competition feature, and a specific feature value of one performance feature can be determined according to the target game performance data matched with the competition feature.

[0043] It should be noted that the specific feature meanings represented by the performance features can be different in different types of competitions, and the specific feature meanings represented by the performance features and the number of specific performance features of each contestant in the current competition are not limited in the embodiments of the present disclosure.

[0044] For example, when the current competition is a game match in a game, the contestants can represent each player participating in the game match, the game performance data can represent game performance data of the player in the game match, and a plurality of competition features that need to be referred to when selecting an MVP (i.e., the target object) in the game match can include the cumulative number of times of killing enemy game characters, the cumulative game damage caused to the enemy game team, etc., and the plurality of performance features of each player in the game match can include the cumulative number of times of killing enemy game characters of each player in the game match, the cumulative game damage caused to the enemy game team, etc.

[0045] For example, when the current game is an online or offline basketball game, the participating objects can represent each basketball player participating in the current game, the game performance data can represent the game performance data of the basketball player in the current game, and the multiple game features that need to be referred to when selecting the MVP (i.e., the selection target object) in the current game can include the cumulative score, the cumulative number of assists, the cumulative number of rebounds, the cumulative number of steals, the cumulative number of blocks, and the like. The multiple performance features of each basketball player in the current game can include the cumulative score, the cumulative number of assists, the cumulative number of rebounds, the cumulative number of steals, and the cumulative number of blocks of each basketball player in the current game.

[0046] Here, in step S101, the game result prediction model represents a classification model pre-trained according to game data of historical games; wherein, taking the number of participating teams in each game as an example, the game performance data of each participating object in a game (i.e., the game data of a game) is input into the game result prediction model, the game result prediction model is used to predict the game result of the game, and the obtained game prediction result can be the probability of the victory of the first participating team (also equivalent to the probability of the failure of the second participating team) and the probability of the failure of the first participating team (also equivalent to the probability of the victory of the second participating team).

[0047] In the embodiments of the present disclosure, according to the above description of the game result prediction model, when the game result prediction model is used to predict the game result of the current game, all performance features of all participating objects in the current game (i.e., the game performance data of all participating objects in the current game) should be used as the model input data of the game result prediction model, that is, the game prediction result output by the game result prediction model can be explained as the result of the joint action of each performance feature of each participating object in the current game. Based on this, when step S101 is performed, the contribution of each performance feature of each participating object in the current game to the game prediction result can be determined by the difference between the game prediction results obtained when the performance feature is included in the model input data and when the performance feature is not included in the model input data.

[0048] Here, considering that the shapley value is a fair contribution allocation method that satisfies four ideal properties (effectiveness, symmetry, additivity, and nullity), as an optional embodiment, the calculation method of the shapley value can be combined with the game result prediction process of the game result prediction model in the following step a. The contribution of each performance feature of each participating object in the game to the game prediction result is quantitatively represented by calculating the shapley value of each performance feature in the game, and the specific steps are as follows:

[0049] Step a, taking the score function in the game result prediction model as the utility function in the shapley value calculation formula, calculating the shapley value of the performance feature as the contribution of the performance feature to the game prediction result output by the game result prediction model.

[0050] Here, the score function represents the function in the game result prediction model for outputting the game prediction result according to the model input data; wherein the model structure to which the game result prediction model belongs is different, and the score function used in the game result prediction model may also be different. For the specific function type to which the score function belongs, the embodiments of the present disclosure do not make any limitation.

[0051] Specifically, in step a, taking the model input data corresponding to the game result prediction model as an example, the n performance features (equivalent to the sum of the performance features of all participating objects is n) can be calculated according to the following improved shapley value calculation formula to obtain the shapley value of the jth performance feature (i.e. the contribution of the jth performance feature to the game prediction result output by the game result prediction model):

[0052] Wherein, N represents a performance feature set composed of n performance features;

[0053] S represents any subset composed of all performance features in the performance feature set N except the jth performance feature;

[0054] |S| represents the number of performance features in the subset S;

[0055] v represents the score function in the game result prediction model (i.e. the utility function in the shapley value calculation formula);

[0056] v(S∪{j}) represents the game prediction result (which can be the probability of victory of the target participating team in the game, wherein the target participating team can represent any participating team in the game) output by the game result prediction model when the model input data of the game result prediction model is the subset S and the jth performance feature;

[0057] v(S) represents the game prediction result (which can also be the probability of the target participating team winning) output by the game result prediction model when the model input data of the game result prediction model is the subset S;

[0058] φ j (v) represents the shapley value of the jth performance feature (i.e., the contribution degree of the jth performance feature to the game prediction result output by the game result prediction model).

[0059] It should be noted that, for v(S∪{j}) and v(S) in the above shapley value calculation formula, since the model input data originally used by the game result prediction model in the game result prediction is the performance feature set N, when calculating v(S∪{j}) and v(S), the number of performance features contained in S∪{j} and the subset S can be less than the performance feature set N. In order to compensate for the impact of the reduction of the number of performance features in the model input data on the game prediction result output by the model, as an optional embodiment, the missing performance features relative to the performance feature set N in S∪{j} and the subset S can be zero-filled when calculating v(S∪{j}) and v(S); as another optional embodiment, the missing features can also be referred to the training data (such as historical game data) used by the game result prediction model in the model training stage, and the feature values of the missing features in the above training data are substituted into the calculation process of v(S∪{j}) and v(S), and the mean value of v(S∪{j}) and the mean value of v(S) calculated multiple times are used as the final v(S∪{j}) and v(S).

[0060] S102, the contribution degrees of the plurality of performance features corresponding to the plurality of performance features are summarized to obtain the contribution degree evaluation result of each participating object in the game.

[0061] Here, referring to the shapley value calculation method shown in step a of the foregoing step S101, the shapley value corresponding to each performance feature (i.e., the contribution degree of each performance feature to the game prediction result output by the game result prediction model) can be calculated. Since each performance feature corresponds to a specific participating object, the shapley values corresponding to all performance features of the same participating object can be summarized, and the contribution degree evaluation result of the participating object in the game can be obtained.

[0062] It should be noted that, referring to the above-mentioned shapley value calculation formula, the relevant description of the game prediction result, when the shapley values corresponding to the performance characteristics of a participating object are summarized, if v(S∪{j}) and v(S) represent the probability that the participating team of the participating object wins, the higher the contribution degree of the performance characteristics of the participating object to the game prediction result indicates the higher the contribution of the participating object to the participating team it belongs to, and the sum of the shapley values corresponding to all performance characteristics of the participating object can be directly summed to obtain the contribution degree evaluation result of the participating object in the game; if v(S∪{j}) and v(S) represent the probability that the participating team of the opponent of the participating object wins, the higher the contribution degree of the performance characteristics of the participating object to the game prediction result indicates the lower the contribution of the participating object to the participating team it belongs to, and the sum of the shapley values corresponding to all performance characteristics of the participating object needs to be summed to obtain the contribution degree evaluation result of the participating object in the game.

[0063] In addition, as another optional embodiment, in order to further improve the accuracy of the above-mentioned contribution degree evaluation result of the same participating object, taking two participating teams in the game (denoted as the first participating team and the second participating team) as an example, the performance characteristics of the participating objects belonging to the first participating team can be sorted in front, and the performance characteristics of the participating objects belonging to the second participating team can be sorted in back in the game data of the game, and the sorted game data is input into the game result prediction model. According to the probability that the game prediction result output by the model is the victory of the first participating team, the sum (denoted as M1) of the shapley values of the performance characteristics of a participating object (denoted as participating object m) is calculated. Then, in the above-mentioned game data, the performance characteristics of the participating objects belonging to the first participating team are sorted in back, and the performance characteristics of the participating objects belonging to the second participating team are sorted in front, and the sorted game data is input into the game result prediction model. According to the probability that the game prediction result output by the model is the victory of the second participating team, the sum (denoted as M2) of the shapley values of the performance characteristics of a participating object (denoted as participating object m) is calculated. When the participating object m belongs to the first participating team, the contribution degree evaluation result of the participating object m in the game can be obtained by M1-M2 (equivalent to when the game prediction result is the probability that the second participating team of the opponent wins, the higher the contribution degree of the performance characteristics of the participating object m to the game prediction result indicates the lower the contribution of the participating object m to the first participating team it belongs to), and when the participating object m belongs to the second participating team, the contribution degree evaluation result of the participating object m in the game can be obtained by M2-M1.

[0064] Specifically, in combination with the optional embodiments described above, in the embodiments of the present disclosure, for a participating object, when the competition prediction result output by the model indicates the probability of the participating team to which the participating object belongs winning, the higher the contribution degree of the performance feature of the participating object to the competition prediction result indicates the higher the contribution of the participating object to the participating team to which the participating object belongs; when the competition prediction result output by the model indicates the probability of the participating team of the opponent of the participating object winning, the higher the contribution degree of the performance feature of the participating object to the competition prediction result indicates the lower the contribution of the participating object to the participating team to which the participating object belongs (i.e., at this time, the shapley value of the plurality of performance features of the participating object should take its opposite number when being summarized).

[0065] S103, determining a target object of the current competition from the participating objects according to the contribution degree evaluation result.

[0066] Here, the greater the numerical value of the contribution degree evaluation result indicates the greater the contribution degree of the performance (i.e., the performance feature described above) of the participating object in the current competition to the competition prediction result output by the competition result prediction model, that is, the greater the numerical value of the contribution degree evaluation result indicates the more critical the role of the participating object in the current competition for the winning team to win (i.e., the competition result), and therefore, the participating object with the largest numerical value of the contribution degree evaluation result (or the participating object with the highest ranking of the contribution degree evaluation result) can be determined as the target object of the current competition (such as the MVP of the current competition).

[0067] Specifically, as an optional embodiment, if the target object selection rule of the current competition does not limit that the target object must come from the winning team (equivalent to when determining the target object, the specific winning team of the current competition is not considered), the participating objects participating in the current competition can be sorted in descending order of the contribution degree evaluation result, and the participating object with the highest ranking of the contribution degree evaluation result is determined as the target object of the current competition.

[0068] Specifically, as an optional embodiment, if the target object selection rule of the current competition requires that the target object be selected from the winning team (such as the competition requiring that the MVP player be selected from the winning team), after the step S102 is executed, the participating objects in the winning team can be sorted in descending order of the contribution degree evaluation result, and the participating object with the highest ranking of the contribution degree evaluation result in the winning team is determined as the target object of the current competition.

[0069] The specific implementation process of each step in the embodiments of the present disclosure will be described in detail as follows:

[0070] For the competition result prediction model in step S101, FIG. 2 shows a flowchart of a method for training the competition result prediction model according to an embodiment of the present disclosure. As shown in FIG. 2, before step S101 is performed, the method includes steps S201-S202, which are specifically as follows:

[0071] S201, inputting the competition data of multiple competitions into the competition result prediction model, and outputting the competition prediction result of each competition by the competition result prediction model.

[0072] Here, the competition data of each competition includes multiple historical performance features of multiple historical participating objects in the competition; where the specific explanation of the competition data and the performance features can refer to the related description in the aforementioned step S101, and the repeated parts will not be described here.

[0073] Here, the competition result prediction model can be a LightGBM model, or a classification model of other model structure. The specific model structure of the competition result prediction model is not limited in the present disclosure.

[0074] It should be noted that since the present disclosure combines the calculation method of shapley value with the competition result prediction process of the competition result prediction model, considering that the treeshap algorithm can be used to quickly calculate the shapley value in the tree structure competition result prediction model, in the present disclosure, the tree structure competition result prediction model (such as the aforementioned LightGBM model) can be preferred.

[0075] S202, training the competition result prediction model according to the loss between the competition prediction result and the real competition result of the same competition, until the competition result prediction model converges.

[0076] Specifically, for each competition, the competition result prediction model can be trained according to the classification loss between the competition prediction result output by the competition result prediction model and the real competition result of the competition, until the competition result prediction model converges, and the competition result prediction model including the adjusted model parameters is obtained.

[0077] With reference to the screening criteria of the performance characteristics in step S101, according to the related description in step S101, in the execution of step S101, according to the various competition characteristics that need to be referred to in the selection of the target object, the target competition performance data matched with each competition characteristic can be extracted from the competition performance data of the contestant as the performance characteristics of the contestant in the competition, wherein, in order to optimize the various competition characteristics that need to be referred to in the selection of the target object, FIG. 3 shows a flowchart of a method for optimizing the various competition characteristics that need to be referred to in the selection of the target object, as shown in FIG. 3, before step S101, the method includes steps S301-S304, specifically:

[0078] S301, taking the score function in the competition result prediction model as the utility function in the shapley value calculation formula, calculating the shapley value corresponding to each competition characteristic in the same historical competition.

[0079] Here, the specific calculation method of the shapley value corresponding to each competition characteristic in step S301 can refer to the specific calculation method of the shapley value corresponding to each performance characteristic in the aforementioned step a, and the repeated parts will not be repeated here.

[0080] S302, according to the shapley value corresponding to the same competition characteristic in the multiple historical competitions, calculating the average shapley value of the competition characteristic.

[0081] Specifically, taking the competition characteristic y1 as an example, if there are m historical competitions, then according to the m shapley values corresponding to the competition characteristic y1 in the m historical competitions, the average shapley value of the competition characteristic y1 in the m historical competitions (i.e. the average of the m shapley values) can be calculated.

[0082] S303, according to the average shapley value of the various competition characteristics, deleting the competition characteristics whose average shapley value meets the target screening condition from the various competition characteristics, and obtaining the remaining target competition characteristics.

[0083] Here, as an optional embodiment, the competition characteristics that meet the target screening condition can be the competition characteristics whose average shapley value is greater than or equal to a preset threshold.

[0084] Here, as another optional embodiment, the game feature meeting the target screening condition can also be: the average shapley value is greater than any other game feature and the difference between the average shapley value and any other game feature is greater than or equal to a preset difference threshold (equivalent to the game feature with the average shapley value far greater than other game features in the above-mentioned multiple game features).

[0085] It should be noted that, with reference to the above two optional embodiments, the game feature to be deleted (i.e., the game feature with the average shapley value meeting the target screening condition) is a game feature with a significantly large average shapley value, i.e., the calculated average shapley value can represent the importance value of a game feature in the above-mentioned multiple game features. When the importance value of a game feature is much greater than other game features, the game feature based on the game feature can cause label leakage, which can reduce the overall calculation accuracy of the shapley value corresponding to the multiple game features. The game feature can be deleted to maintain the overall calculation accuracy of the shapley value corresponding to the multiple game features.

[0086] S304, from the performance data of each of the contestants in the game, determine the performance data matching the target game feature as the multiple performance features of the contestant in the game.

[0087] Here, the specific implementation of step S304 can refer to the above-mentioned related description of game features and performance features in step S101, and the repeated parts will not be described here.

[0088] For example, if the target game feature includes the number of rebounds obtained, then when performing step S304, the number of rebounds obtained by each contestant in the game can be determined from the performance data of each contestant in the game as a performance feature of each contestant in the game.

[0089] Based on the target object determination method shown in steps S101-S103, according to the contribution degree evaluation result of each contestant in each game, the embodiments of the present disclosure further provide the following three optional implementation manners for determining the target object of the season (such as determining the MVP of the season) at the end of the season:

[0090] In one optional embodiment, the target object of the season can be determined in the following steps b1-b2, specifically:

[0091] Step b1, at the end of the season, obtain the average contribution degree evaluation result ranking of the contestants in each winning game in the season as the first average contribution degree evaluation result ranking.

[0092] Here, in step b1, the contribution degree evaluation result ranking of each contestant in each game of the season can be obtained according to the target object determination method shown in steps S101-S103, and then for each contestant, the contribution degree evaluation result rankings of the contestant in the winning games can be filtered out from the obtained contribution degree evaluation result rankings in each game, and the average of the above-mentioned contribution degree evaluation result rankings is calculated as the first average contribution degree evaluation result ranking of the contestant.

[0093] Step b2, the contestant with the highest first average contribution degree evaluation result ranking among the contestants of the season is determined as the target object of the season.

[0094] Here, the first average contribution degree evaluation result ranking represents the average contribution degree evaluation result ranking of a contestant in each winning game of the season, that is, the earlier the first average contribution degree evaluation result ranking (i.e. the smaller the value) is, the greater the contribution of the contestant to the game results of each winning game is, and the more qualified the contestant is to become the target object of the season (such as the MVP of the season).

[0095] In an alternative embodiment, the target object of the season can be determined according to the following steps c1-c2, specifically:

[0096] Step c1, at the end of the season, the average contribution degree evaluation result ranking of the contestant in each game of the season is obtained as the second average contribution degree evaluation result ranking.

[0097] Here, the specific implementation of step c1 can refer to the specific implementation of the aforementioned step b1, except that the winning games in step b1 are replaced by all participating games, and the repeated parts are not described here.

[0098] Step c2, the contestant with the highest second average contribution degree evaluation result ranking among the contestants of the season is determined as the target object of the season.

[0099] Here, the second average contribution degree evaluation result ranking represents the average contribution degree evaluation result ranking of a contestant in each game of the season, that is, the earlier the second average contribution degree evaluation result ranking (i.e. the smaller the value) is, the greater the contribution of the contestant to the game results of each game is, and the more qualified the contestant is to become the target object of the season (such as the MVP of the season).

[0100] In an alternative embodiment, the target object of the season can be determined according to the following steps d1-d2, specifically:

[0101] Step d1, at the end of the season, obtain the average contribution evaluation result of each participating object in each game of the season.

[0102] Here, in step d1, the contribution evaluation result of each participating object in each game of the season can be obtained according to the target object determination method shown in steps S101-S103, and then for each participating object, the average value of the above-mentioned contribution evaluation result is calculated according to the number of games participated by the participating object as the average contribution evaluation result of the participating object in each game of the season.

[0103] Step d2, from the participating objects of the season, determine the participating object with the highest average contribution evaluation result as the target object of the season.

[0104] Here, the average contribution evaluation result represents the average contribution of a participating object in each game of the season, that is, the greater the value of the average contribution evaluation result, the greater the contribution of the participating object to the game results of each game, and the more qualified the participating object is to become the target object of the season (such as the MVP of the season).

[0105] Based on the above-mentioned target object determination method provided by the embodiments of the present disclosure, the contribution degree of the performance feature to the game prediction result output by the game result prediction model is determined under the condition that the performance feature is contained in the model input data of the game result prediction model according to the multiple performance features of each participating object in the game; the contribution degrees corresponding to the multiple performance features are summarized to obtain the contribution evaluation result of each participating object in the game; and the target object of the game is determined from the participating objects according to the contribution evaluation result. In this way, the present disclosure provides a more fair, objective and reasonable basis for the selection of the target object, so that the selection result of the target object has the characteristics of simple transparency and strong interpretability, which is beneficial to help various event organizers more comprehensively and objectively evaluate the performance of each participating object in the game.

[0106] Based on the same inventive concept, the present disclosure also provides a device corresponding to the above-mentioned target object determination method. Since the principle of solving the problem of the target object determination device in the embodiments of the present disclosure is similar to that of the above-mentioned target object determination method in the embodiments of the present disclosure, the implementation of the target object determination device can be referred to the implementation of the above-mentioned target object determination method, and the repeated parts will not be described again.

[0107] Referring to FIG. 4, FIG. 4 shows a structural schematic diagram of a target object determination device provided by an embodiment of the present disclosure, wherein the determination device comprises:

[0108] The contribution allocation module 401 is configured to determine, according to a plurality of performance characteristics of each contestant in the current competition, a contribution degree of each performance characteristic to a competition prediction result output by a competition result prediction model under a condition that the performance characteristic is contained in model input data of the competition result prediction model.

[0109] The statistics module 402 is configured to aggregate the contribution degrees corresponding to the plurality of performance characteristics respectively to obtain a contribution degree evaluation result of each contestant in the current competition.

[0110] The selection module 403 is configured to determine a target object of the current competition from the contestants according to the contribution degree evaluation result.

[0111] In an optional implementation, when the contribution allocation module 401 is configured to determine, according to a plurality of performance characteristics of each contestant in the current competition, a contribution degree of each performance characteristic to a competition prediction result output by a competition result prediction model under a condition that the performance characteristic is contained in model input data of the competition result prediction model, the contribution allocation module 401 is configured to:

[0112] calculate a shapley value of the performance characteristic as the contribution degree of the performance characteristic to the competition prediction result output by the competition result prediction model, by taking a score function in the competition result prediction model as an utility function in a shapley value calculation formula; wherein the score function represents a function in the competition result prediction model for outputting the competition prediction result according to the model input data.

[0113] In an optional implementation, the apparatus further includes a feature screening module, wherein the feature screening module is configured to:

[0114] calculate shapley values corresponding to a plurality of competition characteristics in historical competitions of the same kind by taking a score function in the competition result prediction model as an utility function in a shapley value calculation formula.

[0115] calculate an average shapley value of a competition characteristic according to shapley values corresponding to the competition characteristic in a plurality of historical competitions of the same kind.

[0116] delete, from the plurality of competition characteristics, a competition characteristic whose average shapley value meets a target screening condition, to obtain remaining target competition characteristics according to the average shapley values of the plurality of competition characteristics.

[0117] determine, from performance data of each contestant in the current competition, performance data matched with the target competition characteristics as the plurality of performance characteristics of each contestant in the current competition.

[0118] In an optional implementation, the apparatus further comprises a model training module, wherein the model training module is configured to pre-train the game result prediction model by the following method:

[0119] inputting game data of multiple games into the game result prediction model, and outputting game prediction results of each game by the game result prediction model; wherein the game data comprises multiple historical performance features of multiple historical participating objects in the game;

[0120] training the game result prediction model according to a loss between the game prediction result of the same game and the real game result until the game result prediction model converges.

[0121] In an optional implementation, when the target object of the game is determined from the participating objects according to the contribution degree evaluation result, the selection module 403 is configured to:

[0122] ranking the participating objects of the game according to the contribution degree evaluation result from high to low, and determining the participating object with the highest ranking of the contribution degree evaluation result as the target object of the game.

[0123] In an optional implementation, the apparatus further comprises a season settlement module, wherein the season settlement module is configured to:

[0124] at the time of season settlement, obtaining an average contribution degree evaluation result ranking of the participating object in each winning game of the season as a first average contribution degree evaluation result ranking;

[0125] determining the participating object with the highest ranking of the first average contribution degree evaluation result ranking from the participating objects of the season as the target object of the season.

[0126] In an optional implementation, the season settlement module is further configured to:

[0127] at the time of season settlement, obtaining an average contribution degree evaluation result ranking of the participating object in each game of the season as a second average contribution degree evaluation result ranking;

[0128] determining the participating object with the highest ranking of the second average contribution degree evaluation result ranking from the participating objects of the season as the target object of the season.

[0129] In an optional implementation, the season settlement module is further configured to:

[0130] at the time of season settlement, obtaining an average contribution degree evaluation result of the participating object in each game of the season;

[0131] From the participating objects of the current season, determine the participating object with the highest average contribution evaluation result as the target object of the current season.

[0132] Based on the above-mentioned target object determination apparatus provided by the embodiments of the present disclosure, the contribution degree of each performance feature to the competition prediction result output by the competition result prediction model is determined under the condition that the performance feature is contained in the model input data of the competition result prediction model according to the multiple performance features of each participating object in the current competition; the contribution degrees corresponding to the multiple performance features are summarized to obtain the contribution degree evaluation result of each participating object in the current competition; and the target object of the current competition is determined from the participating objects according to the contribution degree evaluation result. In this way, the present disclosure provides a more fair, objective and reasonable basis for the selection of the target object, so that the selection result of the target object has the characteristics of simplicity, transparency and strong interpretability, which is beneficial to help various event organizers more comprehensively and objectively evaluate the performance of each participating object in the competition.

[0133] Based on the same inventive concept, the present disclosure also provides an electronic device corresponding to the above-mentioned target object determination method. Since the principle of solving the problem in the electronic device in the embodiments of the present disclosure is similar to that of the above-mentioned target object determination method in the embodiments of the present disclosure, the implementation of the electronic device can be referred to the implementation of the above-mentioned target object determination method, and the repeated parts will not be described again.

[0134] FIG. 5 is a structural schematic diagram of an electronic device 500 provided by an embodiment of the present disclosure, which includes a processor 501, a memory 502 and a bus 503. The memory 502 stores machine readable instructions executable by the processor 501. When the electronic device runs a target object determination method as in an embodiment, the processor 501 and the memory 502 communicate through the bus 503. The processor 501 executes the machine readable instructions. When the processor 501 executes the machine readable instructions, the following steps are implemented, specifically:

[0135] According to the multiple performance features of each participating object in the current competition, the contribution degree of each performance feature to the competition prediction result output by the competition result prediction model is determined under the condition that the performance feature is contained in the model input data of the competition result prediction model;

[0136] The contribution degrees corresponding to the multiple performance features are summarized to obtain the contribution degree evaluation result of each participating object in the current competition;

[0137] The target object of the current competition is determined from the participating objects according to the contribution degree evaluation result.

[0138] In an optional implementation, before determining the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model according to the multiple performance features of each contestant in the current competition and respectively determining that the performance features are included in the model input data of the competition result prediction model, the processor 501 is configured to:

[0139] calculate the shapley value of the performance feature as the contribution degree of the performance feature to the competition prediction result output by the competition result prediction model by taking the score function in the competition result prediction model as the utility function in the shapley value calculation formula; wherein the score function represents a function in the competition result prediction model for outputting the competition prediction result according to the model input data.

[0140] In an optional implementation, before determining the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model according to the multiple performance features of each contestant in the current competition and respectively determining that the performance features are included in the model input data of the competition result prediction model, the processor 501 is configured to:

[0141] calculate the shapley value corresponding to each of the multiple competition features in the same historical competition by taking the score function in the competition result prediction model as the utility function in the shapley value calculation formula.

[0142] calculate the average shapley value of the competition feature according to the shapley values corresponding to the competition feature in multiple historical competitions;

[0143] delete the competition features whose average shapley values meet a target screening condition from the multiple competition features according to the average shapley values of the multiple competition features, to obtain target competition features;

[0144] determine the performance data matched with the target competition features from the performance data of each contestant in the current competition as the multiple performance features of each contestant in the current competition.

[0145] In an optional implementation, the processor 501 is configured to obtain the competition result prediction model by pre-training in the following manner:

[0146] input the competition data of multiple competitions into the competition result prediction model, and output the competition prediction results of each competition by the competition result prediction model; wherein the competition data includes multiple historical performance features of multiple historical contestants in the competitions;

[0147] The game result prediction model is trained according to a loss between a game prediction result of the same game and an actual game result until the game result prediction model reaches convergence.

[0148] In an optional implementation, when the target object of the current game is determined from the participating objects according to the contribution evaluation result, the processor 501 is configured to:

[0149] The participating objects of the current game are sorted according to the contribution evaluation result from high to low, and the participating object with the highest contribution evaluation result ranking is determined as the target object of the current game.

[0150] In an optional implementation, the processor 501 is further configured to:

[0151] At the end of the season, an average contribution evaluation result ranking of the participating object in each winning game of the season is obtained as a first average contribution evaluation result ranking;

[0152] From the participating objects of the season, the participating object with the highest first average contribution evaluation result ranking is determined as the target object of the season.

[0153] In an optional implementation, the processor 501 is further configured to:

[0154] At the end of the season, an average contribution evaluation result ranking of the participating object in each game of the season is obtained as a second average contribution evaluation result ranking;

[0155] From the participating objects of the season, the participating object with the highest second average contribution evaluation result ranking is determined as the target object of the season.

[0156] In an optional implementation, the processor 501 is further configured to:

[0157] At the end of the season, an average contribution evaluation result of the participating object in each game of the season is obtained;

[0158] From the participating objects of the season, the participating object with the highest average contribution evaluation result is determined as the target object of the season.

[0159] According to the electronic device provided by the embodiment of the present disclosure, the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model is determined under the condition that the performance feature is contained in the model input data of the competition result prediction model according to the multiple performance features of each contestant in the current competition; the contribution degrees corresponding to the multiple performance features are summarized to obtain the contribution degree evaluation result of each contestant in the current competition; and the target object of the current competition is determined from the contestants according to the contribution degree evaluation result. In this way, the present disclosure provides a more fair, objective and reasonable basis for the selection of the target object, so that the selection result of the target object has the characteristics of simple transparency and strong interpretability, which is beneficial to help various event organizers more comprehensively and objectively evaluate the performance of each contestant in the competition.

[0160] Based on the same inventive concept, the embodiment of the present disclosure also provides a computer readable storage medium, which stores a computer program. When the computer program is run by a processor, the processor performs the following steps:

[0161] According to the multiple performance features of each contestant in the current competition, the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model is determined under the condition that the performance feature is contained in the model input data of the competition result prediction model.

[0162] The contribution degrees corresponding to the multiple performance features are summarized to obtain the contribution degree evaluation result of each contestant in the current competition.

[0163] The target object of the current competition is determined from the contestants according to the contribution degree evaluation result.

[0164] In an optional implementation, in the step of determining the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model under the condition that the performance feature is contained in the model input data of the competition result prediction model according to the multiple performance features of each contestant in the current competition, the processor is configured to:

[0165] The score function in the competition result prediction model is taken as the utility function in the shapley value calculation formula, and the shapley value of the performance feature is calculated as the contribution degree of the performance feature to the competition prediction result output by the competition result prediction model; wherein the score function represents a function in the competition result prediction model for outputting the competition prediction result according to the model input data.

[0166] In an optional implementation, before determining the contribution degree of each performance feature of each contestant in the current competition to the competition prediction result output by the competition result prediction model according to the performance feature included in the model input data of the competition result prediction model, the processor is configured to:

[0167] calculate the shapley value corresponding to each competition feature in the same historical competition as the score function in the competition result prediction model in the utility function in the shapley value calculation formula;

[0168] calculate the average shapley value of the competition feature according to the shapley value corresponding to the competition feature in multiple historical competitions;

[0169] According to the average shapley value of the multiple competition features, delete the competition features whose average shapley value meets the target screening condition from the multiple competition features, and obtain the remaining target competition features;

[0170] From the performance data of each contestant in the current competition, determine the performance data matched with the target competition feature as the multiple performance features of the contestant in the current competition.

[0171] In an optional implementation, the processor is configured to obtain the competition result prediction model by the following method:

[0172] Input the competition data of multiple competitions into the competition result prediction model, and output the competition prediction result of each competition through the competition result prediction model; wherein the competition data includes multiple historical performance features of multiple historical contestants in the competition;

[0173] According to the loss between the competition prediction result and the real competition result of the same competition, the competition result prediction model is trained until the competition result prediction model converges.

[0174] In an optional implementation, when the target object of the current competition is determined from the contestants according to the contribution degree evaluation result, the processor is configured to:

[0175] According to the order from high to low of the contribution degree evaluation result, the contestants participating in the current competition are sorted, and the contestant with the highest ranking in the contribution degree evaluation result is determined as the target object of the current competition.

[0176] In an optional implementation, the processor is further configured to:

[0177] At the end of the season, the average contribution evaluation result ranking of the participating object in each winning game in the season is obtained as a first average contribution evaluation result ranking;

[0178] From the participating objects in the season, the participating object with the highest first average contribution evaluation result ranking is determined as the target object in the season.

[0179] In an optional implementation, the processor is further configured to:

[0180] At the end of the season, the average contribution evaluation result ranking of the participating object in each game in the season is obtained as a second average contribution evaluation result ranking;

[0181] From the participating objects in the season, the participating object with the highest second average contribution evaluation result ranking is determined as the target object in the season.

[0182] In an optional implementation, the processor is further configured to:

[0183] At the end of the season, the average contribution evaluation result of the participating object in each game in the season is obtained;

[0184] From the participating objects in the season, the participating object with the highest average contribution evaluation result is determined as the target object in the season.

[0185] The computer readable storage medium provided by the embodiments of the present disclosure can determine the contribution degree of each performance feature to the game prediction result output by the game result prediction model under the condition that the performance feature is included in the model input data of the game result prediction model according to the plurality of performance features of each participating object in the game; the contribution degrees corresponding to the plurality of performance features are summarized to obtain the contribution evaluation result of each participating object in the game; and the target object of the game is determined from the participating objects according to the contribution evaluation result. In this way, the present disclosure provides a more fair, objective and reasonable basis for the selection of the target object, so that the selection result of the target object has the characteristics of simple transparency and strong interpretability, which is beneficial to help various event organizers to more comprehensively and objectively evaluate the game performance of each participating object.

[0186] In the embodiments of the present disclosure, the computer readable storage medium can also execute other machine readable instructions when run by the processor to perform the target object determination method as described in other embodiments. For specific steps and principles of the target object determination method, refer to the description of the method side embodiments, which will not be repeated here.

[0187] In the embodiments provided by the present disclosure, it should be understood that the disclosed system and method can be implemented in other manners. The embodiments described above are merely exemplary, for example, the division of the units is only a logical function division, and there can be another division manner in actual implementation; for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, or the among different units, can be indirect couplings or communication connections through some interfaces, and electrical, mechanical or other forms.

[0188] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0189] In addition, each functional unit in the embodiments provided by the present disclosure can be integrated into one processing unit, or each unit can exist physically as a separate unit, or two or more units can be integrated into one unit.

[0190] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present disclosure, essentially or part of the technical solutions that make contributions to the prior art, or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present disclosure. The foregoing storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and various media that can store program codes.

[0191] It should be noted that: similar reference numerals and letters in the following drawings represent similar items, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. In addition, the terms "first", "second", "third" and the like are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0192] Finally, it should be noted that the above-described embodiments are merely specific embodiments of the present disclosure, used to illustrate the technical solutions of the present disclosure, and are not intended to limit the present disclosure. The protection scope of the present disclosure is not limited thereto. Although the present disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can make modifications or easy changes to the technical solutions described in the foregoing embodiments, or easily think of changes or equivalent replacements for some of the technical features; and these modifications, changes or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure. All should be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A method for determining a target object, the method comprising: determining, for each contestant object, a contribution degree of each performance feature of the contestant object in a current competition, based on the performance feature, to a competition prediction result output by a competition result prediction model, under a condition that the performance feature is included in model input data of the competition result prediction model; aggregating the contribution degrees corresponding to the performance features to obtain a contribution degree evaluation result of each contestant object in the current competition; and determining a target object of the current competition from the contestant objects based on the contribution degree evaluation result. The determining, for each contestant object, a contribution degree of each performance feature of the contestant object in a current competition, based on the performance feature, to a competition prediction result output by a competition result prediction model, under a condition that the performance feature is included in model input data of the competition result prediction model, comprises: calculating a shapley value of the performance feature as the contribution degree of the performance feature to the competition prediction result output by the competition result prediction model, by taking a score function in the competition result prediction model as an utility function in a shapley value calculation formula, wherein the score function represents a function in the competition result prediction model for outputting the competition prediction result based on the model input data. Before the determining, for each contestant object, a contribution degree of each performance feature of the contestant object in a current competition, based on the performance feature, to a competition prediction result output by a competition result prediction model, under a condition that the performance feature is included in model input data of the competition result prediction model, the method further comprises: calculating shapley values corresponding to multiple competition features in historical competitions of the same kind, by taking a score function in the competition result prediction model as an utility function in a shapley value calculation formula; calculating an average shapley value of a competition feature of the same kind, based on the shapley values corresponding to the competition feature of the same kind in multiple historical competitions; deleting, from the multiple competition features, a competition feature whose average shapley value meets a target screening condition, to obtain target competition features; and determining, from performance data of each contestant object in the current competition, performance data matching the target competition features as the performance features of the contestant object in the current competition. The competition result prediction model is obtained by pre-training in the following manner: inputting competition data of multiple competitions into the competition result prediction model, and outputting competition prediction results of the competitions by the competition result prediction model, wherein the competition data includes multiple historical performance features of multiple historical contestant objects in the competitions; and training the competition result prediction model based on a loss between the competition prediction results and actual competition results of the competitions until the competition result prediction model converges.

2. The determination method according to claim 1, wherein The determining a target object of the current competition from the contestant objects based on the contribution degree evaluation result comprises: ​ 3. The determination method according to claim 2, wherein ​ ​ ​ ​ ​ 4. The determination method according to claim 1, wherein ​ ​ ​ 5. The determination method according to claim 1, wherein ​ According to the contribution degree evaluation results from high to low, the participating subjects are ranked, and the subject with the highest ranking of the contribution degree evaluation results is determined as the target subject of the game.

6. The determination method according to claim 5, wherein The determination method further comprises: At the end of the season, the average contribution degree evaluation result ranking of the participating subjects in each winning game of the season is obtained as a first average contribution degree evaluation result ranking. From the participating subjects of the season, the participating subject with the highest first average contribution degree evaluation result ranking is determined as the target subject of the season.

7. The determination method according to claim 5, wherein The determination method further comprises: At the end of the season, the average contribution degree evaluation result ranking of the participating subjects in each game of the season is obtained as a second average contribution degree evaluation result ranking. From the participating subjects of the season, the participating subject with the highest second average contribution degree evaluation result ranking is determined as the target subject of the season.

8. The determination method according to claim 1, wherein The determination method further comprises: At the end of the season, the average contribution degree evaluation result of the participating subjects in each game of the season is obtained. From the participating subjects of the season, the participating subject with the highest average contribution degree evaluation result is determined as the target subject of the season.

9. A target subject determination device, comprising: a contribution allocation module configured to determine, according to a plurality of performance characteristics of each participating subject in a game, the contribution degree of each performance characteristic to a game prediction result output by a game result prediction model under the condition that the performance characteristic is included in model input data of the game result prediction model; a statistics module configured to aggregate the contribution degrees corresponding to the plurality of performance characteristics respectively to obtain a contribution degree evaluation result of each participating subject in the game; a selection module configured to determine, according to the contribution degree evaluation result, a target subject of the game from the participating subjects.

10. An electronic device comprising: A processor, a memory, and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, the machine readable instructions are executed by the processor to perform the steps of the target subject determination method according to any one of claims 1 to 8.

11. A computer readable storage medium, the computer readable storage medium stores a computer program, the computer program is executed by a processor to perform the steps of the target subject determination method according to any one of claims 1 to 8.

Images