[0041] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.
[0042] In order to solve the above-mentioned technical problem of insufficient scientificity in the virtual prop generation method in the prior art mentioned above, in one embodiment, a method for generating virtual props is specially proposed. This method can be implemented by relying on a computer program. The computer program that runs on a computer system based on the von Neumann system and carries the virtual prop generation method described above may be a virtual prop generation program running on a server, or a virtual prop generation program running on a terminal.
[0043] Specific, such as figure 1 As shown, the method includes the following steps:
[0044] Step S102: Obtain uploaded historical sample data, where the historical sample data includes prop usage parameters and original generation parameters corresponding to the usage parameters.
[0045] In the process of using virtual props, users will generate a series of usage data on the terminal, such as the number of times the props are used, usage scenarios, scores and other data. These data are historical data related to the virtual props, and can be accessed through the Internet The relevant historical data is sent to the server or stored in the memory of the terminal.
[0046] In this embodiment, the server obtains the historical sample data uploaded by the terminal. The so-called historical sample data refers to the data generated during the operation of the terminal of the game user, that is, the original generation of the props used in the terminal during the generation process. Parameters, and the usage parameters of the props under the original generated parameters, for example, the number of times the props are used, the usage scenarios, the user's score, the number of repeated uses of the props, and the target loss rate. It may also be that the terminal obtains relevant data generated by the user during use and stored in the terminal memory. In other words, the execution subject of this method can be a server or a terminal. That is, the process of generating the virtual item can be completed by the server or the terminal. In the case that the generation process of the props is completed by the terminal, the terminal resources can be fully utilized for calculation and analysis, thereby improving the utilization rate of resources.
[0047] It should be noted that what the server obtains can be all historical sample data uploaded by all terminals, historical sample data uploaded by some terminals, or historical sample data uploaded by a certain terminal, and according to the acquired history Depending on the source of the sample data, the items generated by the virtual item generation method will also be different. For example, if the historical data uploaded by a certain terminal is obtained, the generated props are virtual props specific to the user under the terminal. For another example, if the server obtains historical data uploaded by all terminals, the virtual props are based on The virtual props generated by the generation method are better virtual props for all users. If the terminal obtains the historical sample data uploaded by the terminal to the memory, the historical sample data is only the historical data under the terminal, that is, the method for generating the virtual item is the virtual item under the terminal or the user.
[0048] It should be noted that all the historical data of the usage parameters generated during the use of the props are available data, but it can be different according to the actual needs in the subsequent steps. Only certain parts of the parameters are obtained. The acquired parameters can be determined according to specific application scenarios, and when the used parameters are different, the improvement effect or improvement direction of the props produced in the subsequent steps will be different, which can be specifically determined according to user preferences or game needs.
[0049] For example, in an application scenario of a fruit cutting game, for a client of a fruit cutting game running on a terminal, the user will use different props during the game to make the game score higher or easier to pass. The generation of props requires original generation parameters. In this embodiment, the original generation parameters may include but are not limited to the following parameters: the scene where the prop is displayed, the price of the prop, the role of the prop, and so on. The usage parameters of the props include but are not limited to the following parameters: the score obtained by the user cutting fruit in a single round, the score ranking of the user cutting fruit, the fruit loss rate, the user's concentration, the number of replays in a single round, the number of times the prop is used, The effect of the props after use, etc.
[0050] For another example, in an application scenario of a matchmaking game, for a matchmaking game client running on a terminal, the user will use virtual game items such as "refresh" when playing the game. The role of the item In order to rearrange the positions of the targets that need to be eliminated on the game interface, that is, refresh the positions of all the targets. The generation of props requires original generation parameters. In this embodiment, the original generation parameters may include but are not limited to the following parameters: the scene where the prop is displayed, the price of the prop, the role of the prop, and so on. The usage parameters of the props include, but are not limited to, the following parameters: the points obtained by the user in a single game, the scene in which the user uses the prop, the number of times a single game is replayed, the number of times the prop is used, and the effect produced by the prop.
[0051] Step S104: Calculate the evaluation reference value of the virtual item according to the historical sample data, and the evaluation reference value matches the original generation parameter.
[0052] The evaluation reference value of a virtual item is a quantitative value that can be used to evaluate a certain index or the overall usage of the virtual item, and the evaluation reference values of different items can be compared, so as to be in a certain dimension Compare virtual items. For example, the user’s preference value for the props, the purchase rate of the props, the impact value of the use of the props on the score, or other specific values that can be used to quantitatively evaluate an indicator of the props can be used as an evaluation of a virtual prop Reference. Further, the evaluation reference value can be obtained or calculated from the specific data contained in the historical sample data, each set of props under the original generation parameters corresponds to a set of historical sample data, and the corresponding virtual props evaluation reference value can be obtained .
[0053] In this embodiment, the historical sample data for the calculation of the evaluation reference value may be preprocessed before the calculation. For example, de-noising processing of historical sample data can clarify the noise and inconsistencies in the data, improve data quality, make subsequent further processing or analysis of sample data more accurate and consistent, and improve data credibility. For another example, the preprocessing of historical sample data can also include sampling processing, such as random sampling, stratified sampling, cluster sampling, etc., using a small amount of data to replace the original large amount of data for analysis, which can reduce the amount of data processed. And the amount of calculation, especially when the sample data volume is huge, in this embodiment, when the server processes all the historical sample data uploaded by the user terminal, it is necessary to sample the sample data first. Further, the preprocessing of historical sample data may also include other data processing methods, such as data transformation, that is, compressing data with relatively large values or data with a large range of values to a smaller interval to facilitate processing and calculation. In addition, the specific sample data processing can be determined based on the specific circumstances of the historical sample data and the direct or indirect purpose of the subsequent calculation steps to determine whether to use data preprocessing and which data preprocessing method to use.
[0054] It should be noted that the prop usage parameters and original generation parameters included in the historical sample data used in calculating the evaluation reference value are all quantized parameter values. For example, in an application scenario of a fruit cutting game, the parameter "scene displayed by the props" is the number of all scene numbers in the game application, and the parameter "user’s score ranking for fruit cutting" is the user’s score ranking among all users The ratio of the number of all users, the parameter "effects produced after the use of the item" is the ratio of the user's score per unit time within a certain time threshold after using the item to the score per unit time during the entire game. And the specific quantization method and the interval size of the quantization value can be set according to actual needs.
[0055] The following uses specific embodiments to illustrate the calculation method of the evaluation reference value of the virtual prop:
[0056] For example, in an application scenario of a fruit cutting game, in the acquired historical sample data, the original generation parameters include: the display scene of the props, the price of the props, and the role of the props; the usage parameters of the props include: user single-round cut The score obtained by the fruit, the ranking of the user's score for cutting the fruit, the fruit loss rate, the user's concentration, the number of replays in a single game, the number of times the prop is used, and the effect of the prop after use. After the server obtains the above-mentioned parameters, the above-mentioned parameters are quantified according to a preset standard, so as to facilitate the calculation of the following evaluation reference values.
[0057] Let x 1 ,x 2 ,x 3 Respectively represent the quantitative values of the parameters "scene displayed by props", "price of props" and "function of props", y 1 ,y 2 ,...,Y 7 Respectively indicate the parameters "the user's score obtained by cutting fruit in a single round", "the user's score ranking of fruit cutting", "fruit churn rate", "user concentration", "number of replays in a single round", and "use of props" The quantified value of "number of times" and "effects produced after the item is used".
[0058] Remember And t=X+Y, then according to the formula
[0059] f ( t ) = 1 2 π exp ( - t 2 2 )
[0060] Calculate the item purchase rate f(t), which is the evaluation reference value of the virtual item, where a 1 ,a 2 ,a 3 And b 1 ,b 2 ,...,B 7 Are weighting coefficients, and ∑a i =1,a i ∈[0,1], ∑b i =1,b i ∈[0,1].
[0061] It should be noted that for the same historical sample data, there may be different sample data processing, or different evaluation reference values, or different calculation formulas or calculation parameters for the same evaluation reference value.
[0062] Step S106: Determine whether the evaluation reference value meets the preset value, and if the judgment result is yes, obtain the original generation parameter that matches the evaluation reference value and assign it to the target generation parameter; The historical sample data and the evaluation reference value are used to calculate target generation parameters through a preset machine learning algorithm.
[0063] For the evaluation reference value, it indicates to a certain extent the pros and cons of the corresponding virtual item, or to a certain extent the user’s preference for the item, and the evaluation reference value has a corresponding optimal value or maximum value. Expected value. When the evaluation reference value reaches the optimal value or the maximum expected value, it means that the relevant parameter of the item meets the needs of the user or is liked by the user or is generally used by the user. If the evaluation reference value does not reach the optimal value or the maximum expected value, it means that the relevant parameters of the item do not meet the needs of the user or are not liked by the user, and further improvements or upgrades are needed.
[0064] Generally speaking, the preset value is a value or an interval set in advance according to needs, and the judgment rule for evaluating whether the reference value meets the preset value is set according to the actual situation. For example, the preset value is an interval, may wish to be set to [M, +∞], then the judgment rule for whether the evaluation reference value meets the preset value can be judging that the calculated evaluation reference value is greater than or equal to M, that is, judging the evaluation reference The value meets the preset value, that is, the judgment rule is to evaluate whether the reference value belongs to the preset interval. For another example, if the preset value is a specific value N, the judgment rule may be to determine whether the evaluation reference value is equal to the value N, or whether it is greater than the value N. Of course, the judgment rule for evaluating whether the reference value meets the preset value may also be other rules.
[0065] If the step: the judgment result of judging whether the evaluation reference value meets the preset value is yes, it means that the relevant parameter of the prop corresponding to the evaluation reference value meets the preset conditions, no upgrade or improvement is required, and you can continue Use the prop. Therefore, the historical sample data corresponding to the evaluation reference value meeting the preset value can be directly obtained, and the value of the corresponding original generation parameter can be obtained and directly assigned to the target generation parameter.
[0066] It should be noted that if in the above judgment step, if multiple evaluation reference values meet the preset value, the optimal evaluation reference value should be selected from them, and the historical sample data corresponding to the optimal evaluation reference value should be obtained , Get the value of the corresponding original generation parameter and directly assign it to the target generation parameter.
[0067] In another embodiment, in view of the possibility that multiple evaluation reference values all meet the preset value, after the step of determining whether the evaluation reference value meets the preset value, and in the case where the above judgment result is yes, According to the judgment result, all the evaluation reference values that meet the preset value are obtained, and the quantity is obtained. When the number is equal to 1, the historical sample data corresponding to the evaluation reference value that meets the preset value is directly obtained, and the value of the corresponding original generation parameter is obtained and directly assigned to the target generation parameter.
[0068] Correspondingly, when the number is greater than 1, that is, more than one evaluation reference value that meets the preset value, it is necessary to select the optimal evaluation reference value from multiple evaluation reference values that meet the preset value, and then obtain The original generation parameter corresponding to the optimal evaluation reference value is assigned to the target generation parameter. Moreover, the above-mentioned process of selecting an optimal evaluation reference value from a plurality of evaluation reference values satisfying the preset value is a process of searching for the optimal value from a plurality of values according to a preset rule.
[0069] For example, when the preset value of the evaluation reference value is a relatively large interval, the evaluation reference value calculated in the sample data is generally more likely to meet the preset value. In the case of reasonable prop design, The number of evaluation reference values that meet the preset value may be greater than one.
[0070] It should be noted that the standard for evaluating the quality of the evaluation reference value is different according to different specific application scenarios. For example, when the evaluation reference value represents the purchase rate of an item, the larger the evaluation reference value, the higher the purchase rate, that is, the larger the evaluation reference value, the better. Therefore, the process of finding the optimal value is a process of finding the maximum value. For example, if the evaluation reference value is the churn rate of the item, the greater the churn rate, the worse the use of the item, and the smaller the evaluation reference value, the better. Therefore, the process of finding the optimal value is a process of finding the minimum value. process.
[0071] If the judgment result of judging whether the evaluation reference value meets the preset value is no, it indicates that the evaluation of the existing props does not meet the preset standard, and the props corresponding to the corresponding original generation parameters do not meet the requirements of the user , Needs to be improved or optimized. For the optimization or improvement of props, historical sample data can be analyzed to analyze the influence of prop usage parameters and original generated parameters on the evaluation reference value, as well as the mutual influence between the parameters, so as to calculate the evaluation reference that can reach the preset value The value of the target parameter.
[0072] Specifically, the process of analysis and calculation can be completed by machine learning algorithms, and includes an environment module, a learning module, and an execution module. The environment module provides the historical sample data obtained by the system to the learning module for learning processing. The learning module uses the historical sample data provided by the environment module to establish and modify the knowledge base, which is the general principle of the execution module for data processing. Or it can be an action function that acts on sample data. The execution module completes calculation and analysis based on the knowledge base, and feeds relevant information back to the learning module. The learning module continuously modifies and improves the knowledge base based on the feedback information to obtain the best The knowledge base is the relationship between the sample data and the evaluation reference value, so that the value of each parameter when the evaluation reference value meets the preset value can be obtained according to this relationship.
[0073] It should be noted that the calculation method of target generation parameters is not limited to machine learning algorithms. As long as the influence of the parameter on the evaluation reference value can be analyzed to calculate the target generation parameter that enables the evaluation reference value to reach the preset value, it can be used in this implementation. The calculation of the target parameter in the example.
[0074] In one embodiment, the calculation of target generation parameters is obtained by using the BP neural network algorithm as an example, based on historical sample data and calculated evaluation reference values. The specific calculation process is as follows:
[0075] Because the usage parameters are the usage parameters under the virtual props corresponding to the original generation parameters, that is to say, the usage parameters are changed according to the changes of the original generation parameters and are determined by the original generation parameters. However, the original generation parameters The specific process of how to influence usage parameters is unclear. With the original generation parameters as the input layer and the usage parameters as the output layer, the neural network is trained with historical sample data until the output error is less than the preset maximum error threshold. According to the BP neural network model that has been obtained, the original generation parameters can be directly input, and the corresponding usage parameters can be obtained.
[0076] Of course, the calculation of the above parameters can also be obtained by other algorithms, and any algorithm that can calculate the mutual relationship between the parameters can be used, and it is not limited to the example of the BP algorithm cited here.
[0077] It should be noted that the same method can be used to build a BP neural network model with the usage parameters as the input layer and the original generation parameters as the output layer, so that the corresponding original generation parameters can be obtained according to the usage parameters.
[0078] Therefore, the mapping relationship between the output layer and the input layer parameters can be obtained according to the foregoing method, and the mapping relationship can be used to obtain the specific values of another part of the parameters when a certain part of the parameters is determined.
[0079] Further, in order to obtain the mapping relationship between each parameter and the evaluation reference value, a BP neural network model with the original generation parameters and usage parameters as the input layer and the evaluation reference value as the output layer is established. The BP neural network is trained with the parameters contained in the historical sample data and the evaluation reference values calculated in the foregoing steps, so that the mapping relationship or mutual influence between the input layer parameters and the output layer evaluation reference values can be obtained.
[0080] In the case that the evaluation reference value cannot meet the preset value, according to the above-mentioned mapping relationship between the parameters based on the BP neural network model, determine the value of each parameter required when the evaluation reference value meets the preset value The value is the target generation parameter required for the generation of the virtual item.
[0081] Step S108: Send the target generation parameter to the terminal, so that the terminal generates a target virtual item corresponding to the target generation parameter according to the target generation parameter.
[0082] After the target generation parameters are generated, the specific settings of the target virtual props can be determined according to the relevant parameter values of the target generation parameters, that is, the target virtual props corresponding to the target generation parameters are determined. Moreover, the generation of the target virtual item is completed by the terminal according to the target generation parameters.
[0083] In this embodiment, the server may generate target generation parameters, and then send the target generation parameters to the terminal, and then the terminal generates the corresponding target virtual props according to the target generation parameters; it may also be that the terminal generates the target generation parameters, and then the target The generation parameter is sent to the generation module of the virtual prop, and the virtual prop module generates the corresponding target virtual prop according to the target generation parameter.
[0084] It should be noted that the generation of the target virtual item may also be completed by the server according to the target generation parameters, and then the virtual item is sent to the terminal for the user to use at the terminal.
[0085] Such as figure 2 As shown, in order to solve the above-mentioned technical problem of insufficient scientificity in the virtual prop generation method in the prior art mentioned above, in one embodiment, a virtual prop generation device is specially proposed, which includes a historical sample data acquisition module 102 , Evaluation reference value calculation module 104, judgment module 106, target virtual prop generation module 108, specifically,
[0086] The historical sample data obtaining module 102 is configured to obtain historical sample data uploaded by the terminal, where the historical sample data includes prop usage parameters and original generation parameters corresponding to the usage parameters;
[0087] The evaluation reference value calculation module 104 is configured to calculate an evaluation reference value of the virtual item according to the historical sample data, where the evaluation reference value matches the original generation parameter;
[0088] The judging module 106 is configured to judge whether the evaluation reference value meets the preset value, and if not, calculate the target generation parameter through a preset machine learning algorithm according to the historical sample data and the evaluation reference value;
[0089] The target virtual prop generation module 108 is configured to send the target generation parameter to the terminal, so that the terminal generates a target virtual prop corresponding to the target generation parameter according to the target generation parameter.
[0090] Optionally, the judging module 106 is further configured to obtain an original generation parameter matching the evaluation reference value when the evaluation reference value meets a preset value, and assign a value to the target generation parameter.
[0091] Optionally, the judgment module 106 is further configured to calculate the mapping relationship between the original generation parameters and the prop usage parameters through the preset machine learning algorithm, and the original generation parameters and the prop usage parameters. A mapping relationship between a situation parameter and the evaluation reference value; according to the mapping relationship, the target generation parameter corresponding to the evaluation reference value that satisfies the preset value is calculated.
[0092] Optionally, the judgment module 106 is further configured to: obtain the number of the evaluation reference values that satisfy the preset value;
[0093] When the number of the evaluation reference values satisfying the preset value is greater than 1, among the evaluation reference values satisfying the preset value, the optimal evaluation reference value is searched, and the optimal evaluation reference value is obtained. Corresponding original generation parameters, and assigning values to the target generation parameters;
[0094] When the number of the evaluation reference values satisfying the preset value is not greater than 1, the step of obtaining the original generation parameter matching the evaluation reference value and assigning the value to the target generation parameter is performed.
[0095] Optionally, the evaluation reference value calculation module 104 is further configured to preprocess the historical sample data, and the preprocessing method includes noise processing, sampling processing, and/or data transformation processing.
[0096] Implementing the embodiments of the present invention will have the following beneficial effects:
[0097] After adopting the above-mentioned method and device for generating virtual props, when the user uses the props, the relevant data generated during the user's use of the props will be counted, and based on the historical data of the user, the specifics of the virtual props will be obtained through machine learning algorithms. The analysis and evaluation of the use situation takes into account the actual situation of the user when using the props, which is equivalent to that the user can give feedback to the props through the relevant parameters in actual use. If the relevant evaluation of the prop does not meet the expectations, it means that the actual use of the prop does not meet the needs of the user, so timely correction is required. Then through the preset machine learning algorithm and the analysis of historical data, the generation parameters that can make the evaluation of the props reach a better situation are calculated, which is equivalent to improving or upgrading the props to make the props more in line with the needs of users. It is more reasonable and scientific, and the user experience is improved. Further, according to the different historical data used in the evaluation and improvement of the props, for example, if the data used is the historical data of all users, the improvement of the props is aimed at all users, and the generated items are more universal Props; if the relevant data generated by a specific user in the process of using the props is used, the above improvements are aimed at the user, and the generated props are personal props that are more in line with the user’s usage. Ground increases the science of props generation and further enhances the user experience.
[0098] In the description of this specification, descriptions with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean specific features described in conjunction with the embodiment or example , Structure, materials or features are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above-mentioned terms are not necessarily for the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples and the features of the different embodiments or examples described in this specification without mutual contradiction.
[0099] Those skilled in the art can combine and combine the different embodiments and the features of the different embodiments described in this specification.
[0100] The steps in the methods of all embodiments of the present invention can be sequentially adjusted, combined, and deleted according to actual needs; the modules or units in the devices of all embodiments of the present invention can be combined, divided, and deleted according to actual needs.
[0101] The logic and/or steps represented in the flowchart or described in other ways herein, for example, can be considered as a sequenced list of executable instructions for implementing logic functions, and can be embodied in any computer-readable medium, For use by instruction execution systems, devices, or equipment (such as computer-based systems, systems including processors, or other systems that can fetch and execute instructions from instruction execution systems, devices, or equipment), or combine these instruction execution systems, devices Or equipment.
[0102] Those of ordinary skill in the art can understand that all or part of the steps carried in the method of the foregoing embodiments can be implemented by a program instructing relevant hardware to complete, and the program can be stored in a computer-readable storage medium. When executed, it includes one of the steps of the method embodiment or a combination thereof.
[0103] In addition, the functional units in the various embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or software functional modules. If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it may also be stored in a computer readable storage medium.
[0104] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. A person of ordinary skill in the art can comment on the above-mentioned embodiments within the scope of the present invention. The embodiment undergoes changes, modifications, replacements and modifications.
[0105] The above-disclosed are only preferred embodiments of the present invention. Of course, the scope of rights of the present invention cannot be limited by this. Those of ordinary skill in the art can understand all or part of the process of implementing the above-mentioned embodiments and follow the claims of the present invention. The equivalent changes made are still within the scope of the invention.