CN113191810A - Game index prediction method and device and electronic equipment - Google Patents
Game index prediction method and device and electronic equipment Download PDFInfo
- Publication number
- CN113191810A CN113191810A CN202110488896.2A CN202110488896A CN113191810A CN 113191810 A CN113191810 A CN 113191810A CN 202110488896 A CN202110488896 A CN 202110488896A CN 113191810 A CN113191810 A CN 113191810A
- Authority
- CN
- China
- Prior art keywords
- game
- data
- prediction
- model
- stream
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/70—Game security or game management aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Multimedia (AREA)
- Educational Administration (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Computer Security & Cryptography (AREA)
- Evolutionary Computation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a game index prediction method and device and electronic equipment. Wherein, the method comprises the following steps: acquiring a data stream corresponding to the current moment of a game to be predicted; the data stream includes: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes; carrying out game index prediction at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model; and generating game index prediction data of the next moment according to the first prediction result and the second prediction result, so that the real-time prediction according to the data stream generated by the game can be realized through the batch processing model and the stream calculation model, and the prediction accuracy of the game index is improved.
Description
Technical Field
The present invention relates to the field of game technologies, and in particular, to a method and an apparatus for predicting game indexes, and an electronic device.
Background
The game index is a unified data analysis index set in the game operation process according to international specifications, and mainly comprises the following steps: the number of newly increased users, the number of active users, the daily income of games and the like are increased every day, the ratio of promotion input and output of games can be calculated by predicting the daily income of games, and the number of newly increased users can be used for sensing the operation activities planned in the future and the influence of promotion strategies on game promotion and the like by predicting the daily income of games, so that the understanding and analysis of user behaviors such as game planning and game operation are facilitated, and game products are improved.
The existing prediction method mainly obtains historical related data, including game index actual data and game related data influencing game indexes, disperses the historical related data into time sequence data, such as one data or a plurality of fixed data for batch processing every day, trains a prediction model, and deploys online for service until the prediction effect meets the requirement. In the online service process, the prediction model is triggered at regular time to predict, for example, after the latest data is acquired in the morning, the game indexes of the next day or a plurality of days in the future are predicted. Therefore, although the game index can be predicted by the conventional method, historical related data needs to be converted into discrete time data for batch processing, and cannot be updated in time, so that the predicted numerical value has large hysteresis and deviation, and the prediction precision of the game index is influenced.
Disclosure of Invention
In view of the above, the present invention provides a method, an apparatus and an electronic device for predicting game indexes, so as to alleviate the above problems and improve the accuracy of game index prediction.
In a first aspect, an embodiment of the present invention provides a game index prediction method, in which a pre-trained prediction model is provided by an electronic device, the prediction model includes a batch processing model for predicting game indexes for bounded data and a flow calculation model for predicting game indexes for unbounded data, the method includes: acquiring a data stream corresponding to the current moment of a game to be predicted; wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes; carrying out game index prediction at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model; and generating game index prediction data of the next moment according to the first prediction result and the second prediction result.
With reference to the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, where the step of performing game index prediction at a time next to the current time based on the data stream and the prediction model includes: acquiring a bounded data set corresponding to a data stream; wherein the bounded data set carries a start time identifier and an end time identifier; inputting the bounded data set into a batch processing model of a prediction model for prediction processing; the data stream is input to a stream calculation model of the prediction model to perform prediction processing.
With reference to the first possible implementation manner of the first aspect, an embodiment of the present invention provides a second possible implementation manner of the first aspect, where the step of obtaining a bounded data set corresponding to a data stream includes: cutting the data stream according to a preset sliding window to obtain data sub-streams corresponding to a plurality of windows; wherein, the data sub-stream includes the start time mark and the end time mark of the window corresponding to the data sub-stream; a plurality of data sub-streams are grouped into a bounded data set.
With reference to the first aspect, an embodiment of the present invention provides a third possible implementation manner of the first aspect, where the step of generating game index prediction data at a next time according to the first prediction result and the second prediction result includes: and carrying out weighted calculation on the first prediction result and the second prediction result according to the corresponding weight values to obtain game index prediction data at the next moment.
With reference to the third possible implementation manner of the first aspect, an embodiment of the present invention provides a fourth possible implementation manner of the first aspect, where the step of performing weighted calculation on the first predicted result and the second predicted result according to corresponding weight values includes: responding to the weight value adjustment operation aiming at the game to be predicted, and adjusting the weight values corresponding to the first prediction result and the second prediction result respectively; and carrying out weighted calculation on the first prediction result and the second prediction result according to the adjusted weight value.
With reference to any one of the foregoing possible implementation manners of the first aspect, an embodiment of the present invention provides a fifth possible implementation manner of the first aspect, where the step of obtaining a data stream corresponding to a current time of the game to be predicted includes: monitoring a log file of a game to be predicted in real time to obtain game related data which influences game indexes and is generated by the game to be predicted at the current moment; acquiring game index actual data corresponding to the current moment from the service data of the game to be predicted; and forming the game related data and the game index actual data into a data stream corresponding to the current moment.
With reference to the first aspect, an embodiment of the present invention provides a sixth possible implementation manner of the first aspect, where the method further includes: and updating the flow calculation model according to the error between the second prediction result and the actual data of the game index.
With reference to the sixth possible implementation manner of the first aspect, an embodiment of the present invention provides a seventh possible implementation manner of the first aspect, wherein the step of updating the stream computation model according to the error between the second predicted result and the game index actual data includes: acquiring second prediction results and game index actual data which respectively correspond to a specified number of moments before the current moment; for each moment, calculating a difference value between a second prediction result corresponding to the moment and actual data of the game index, and determining an increment corresponding to the current moment based on the difference value corresponding to each moment; and updating the flow calculation model according to the increment corresponding to the current moment.
With reference to the first aspect, an embodiment of the present invention provides an eighth possible implementation manner of the first aspect, where the training process of the prediction model includes: training an original batch processing model by applying a first data stream sample to obtain a trained batch processing model; inputting the second data stream sample into a batch processing model to obtain a first prediction training result corresponding to the second data stream sample; training a raw stream calculation model by using the truth value and the unbounded data stream sample corresponding to the second data stream sample to obtain a trained stream calculation model by using the first prediction training result as a truth value of the second data stream sample; and taking the trained batch processing model and the trained flow calculation model as a prediction model of the game index.
In a second aspect, an embodiment of the present invention further provides a game index prediction apparatus, which provides a pre-trained prediction model through an electronic device, where the prediction model includes a batch model for predicting game indexes for bounded data and a flow calculation model for predicting game indexes for unbounded data, and the apparatus includes: the data flow acquisition module is used for acquiring a data flow corresponding to the current moment of the game to be predicted; wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes; the game index prediction module is used for predicting the game index at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model; and the game index prediction data generation module is used for generating game index prediction data of the next moment according to the first prediction result and the second prediction result.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the game index prediction method according to the first aspect when executing the computer program.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the game index prediction method in the first aspect are performed.
The embodiment of the invention has the following beneficial effects:
the embodiment of the invention provides a game index prediction method, a game index prediction device and electronic equipment, which are used for predicting a game index at the next moment based on a data stream and a prediction model corresponding to the current moment, wherein the prediction model comprises a batch processing model and a stream calculation model.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a method for predicting game indexes according to an embodiment of the present invention;
fig. 2 is a flowchart of an updating method of a stream computation model according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for training a prediction model according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a game index prediction according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating a game index prediction apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the field of games, game planning, game operation and the like can be helped to understand and analyze user behaviors by predicting game indexes so as to improve game products. The existing method mainly converts historical related data into discrete time data for batch processing and predicts at regular time, for example, game indexes of one or more days in the future are predicted at one time, so that the predicted numerical value has larger hysteresis and deviation and cannot be updated in time, and the prediction accuracy of the game indexes is influenced.
Aiming at the technical problem, the embodiment of the invention provides a game index prediction method, a game index prediction device and electronic equipment, wherein a flow calculation model is added in a prediction model, so that data generated by a game to be predicted can be calculated in real time, and the data of a game player can be dynamically refreshed in real time through real-time data calculation and analysis, so that the real-time prediction of the game index is realized, the problems of large hysteresis and errors of a prediction value caused by batch processing in the conventional method are solved, the timeliness and accuracy of the prediction of the game index are improved, and the game index prediction method has high practical value.
To facilitate understanding of the present embodiment, a detailed description is provided below of a game index prediction method according to an embodiment of the present invention. Wherein the executing subject is an electronic device provided with a pre-trained prediction model comprising a batch model for predicting game metrics for bounded data and a flow calculation model for predicting game metrics for unbounded data.
The bounded data and the unbounded data are relative concepts, and are mainly determined according to the time range. The data Processing method for the bounded data set is called Batch Processing, for example, the bounded data is read from a system such as a relational database or a file system and then processed in a distributed system, and finally, the Processing result is written into a storage medium. The existing distributed batch processing framework mainly comprises Apache Hadoop, Apache Spark and the like. The unbounded data in the unbounded data set is data that is continuously generated from the beginning of generation, and the data processing method for the unbounded data set is called Streaming data processing, which is simply referred to as Streaming processing. The stream processing process can realize real-time processing of unbounded data, and a stable and continuous processing flow is kept between data production and consumption. The existing distributed computing engines such as Apache Storm, Spark Streaming, Apache Flink and the like can support processing unbounded data to different degrees.
In order to perform unified data processing on bounded data sets and unbounded data sets, the conventional open-source big data processing framework can simultaneously support stream type calculation and batch calculation, and mainly comprises two frameworks, namely an Apache Spark framework and an Apache flight framework. The Apache Spark processes different types of data sets uniformly through a batch processing mode, and for an unbounded data set, data is divided into micro-batches (bounded data sets) according to batches for processing; apache Flink is used for processing different types of data sets uniformly through a stream processing mode from another angle. In practical applications, users may need to use multiple computing frameworks in parallel to solve different types of data processing, for example, users may use Apache Flink as an engine for stream computing and Apache Spark or MapReduce as an engine for batch computing, which not only increases the complexity of the system, but also increases the cost for user learning and operation and maintenance. Because Apache Flink comparison conforms to a rule mode of data generation, bounded data can be converted into unbounded data to be uniformly streamed, and finally batch processing and stream processing are uniformly performed in a set of streaming engine, the electronic equipment in the embodiment of the invention preferably adopts an Apache Flink framework, so that not only can the prediction tasks of batch calculation and stream calculation be realized, but also the cost is saved.
Based on the electronic device, an embodiment of the present invention provides a game index prediction method, as shown in fig. 1, the method includes the following steps:
step S102, acquiring a data stream corresponding to the current moment of the game to be predicted;
wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes; the game index actual data comprises but not limited to the number of newly increased daily users, the number of active daily users and the daily income of the game, and the game related data comprises but not limited to the number of active daily games and the number of newly increased daily games which affect the daily income of the game, and the game index actual data can be specifically set according to actual conditions.
In practical application, the data stream at the current time can be obtained according to the following process: monitoring a log file of a game to be predicted in real time to obtain game related data which influences game indexes and is generated by the game to be predicted at the current moment; acquiring game index actual data corresponding to the current moment from the service data of the game to be predicted; and forming the game related data and the game index actual data into a data stream corresponding to the current moment. Specifically, the log file of the game to be predicted in the log server can be monitored in real time, and when a new log file is monitored, the new log file is directly acquired from the log server to obtain game related data generated by the game to be predicted at the current moment; and the actual data of the game index at the current moment is obtained from the business data stored in the business database, so that the data stream at the current moment is obtained according to the actual data of the game index at the current moment and the game associated data, the data content in the data stream is enriched, and the prediction accuracy of the game index is improved.
Step S104, performing game index prediction at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model;
the data stream obtained above is actually unbounded, while the batch model can process only bounded data, and therefore, the data stream needs to be converted into a corresponding bounded data set. One possible prediction method includes: acquiring a bounded data set corresponding to a data stream; the bounded data set carries a start time identification and an end time identification; inputting the bounded data set into a batch processing model of a prediction model for prediction processing; the data stream is input to a stream calculation model of the prediction model to perform prediction processing. Specifically, an unbounded data stream is converted into a corresponding bounded data set, so that the bounded data set can be input into a batch processing model for prediction processing to obtain a first prediction result; and inputting the data stream into the stream calculation model for prediction processing according to an event triggering mode to obtain a second prediction result. It should be noted that the event triggering mode is only an expression mode of stream calculation, and in practical application, generation of each log and each event can be used as one event trigger, and can be specifically set according to actual conditions, so that timeliness of game index prediction is guaranteed.
In practical application, the bounded data and unbounded data can be converted into each other, for example, order transaction data stored in the system for one year is a bounded data set in nature, and if the order transaction data are sent to the streaming system one by one according to a generated sequence and processed by the streaming system, the data can be considered to be relatively unbounded, that is, the unbounded data set is obtained at this time; similarly, the unbounded data can be split into bounded data for processing, for example, data generated by the system is accessed into a storage system, the data is cut according to the year or month, so that the data is divided into bounded data sets with different time lengths, and finally the data is processed in a batch processing mode.
Therefore, the bounded data set corresponding to the data flow can be obtained as follows: cutting the data stream according to a preset sliding window to obtain data sub-streams corresponding to a plurality of windows; wherein, the data sub-stream includes the start time mark and the end time mark of the window corresponding to the data sub-stream; a plurality of data sub-streams are grouped into a bounded data set. Specifically, after the data stream is cut according to a preset sliding window, the data substreams corresponding to the multiple windows may be stored through a big data related database or a File storage System in the electronic device, such as HBase, HDFS (Hadoop Distributed File System, Distributed File System), and the like, and then the Distributed computation logic is triggered at regular time, such as a bounded data set is formed by the multiple data substreams by using the technologies of Apache Spark, MapReduce, and the like; or directly combining the multiple data sub-streams into a bounded data set according to the distributed computation logic, and then storing the data sub-streams corresponding to the multiple windows, wherein the specific computation and storage sequence can be set according to the actual situation. In addition, the data stream can be converted into the corresponding bounded data set by acquiring the bounded data set corresponding to the data stream in a manual control manner, for example, a user directly sets the start time and the end time of the bounded data set, or sets the start time and the end time of each data sub-stream, and calculating a plurality of data sub-streams.
In practical application, for a part of games to be predicted with little change of game indexes, the change trend of the game indexes for a longer time such as a future day can also be directly predicted, that is, another possible prediction mode includes: acquiring a bounded data set corresponding to a data stream; the bounded data set carries a start time identification and an end time identification; and inputting the bounded data set into a batch processing model of the prediction model for prediction processing to obtain a first prediction result. Specifically, in the prediction manner, a user or a developer may autonomously select a batch processing model of the prediction model to predict some game indexes, and periodically trigger the batch processing model to predict the game indexes in a longer time in the future to obtain a first prediction result, such as a prediction of a change trend of the game indexes in one or more days in the future at 00:00 a day, where the obtaining manner of the bounded data set corresponding to the data stream may refer to the above embodiment, and details of the embodiment of the present invention are not described here.
And at a certain moment in the prediction process, the prediction can be performed only by the flow calculation model, and another possible prediction mode at this time includes: and inputting the data stream into a stream calculation model of the prediction model for prediction processing to obtain a second prediction result. Specifically, the data stream at the current time may be directly input to the stream calculation model in an event-triggered manner for prediction processing, so as to obtain a second prediction result, where the first prediction result is a prediction result of the batch processing model at a previous time of the current time, that is, a first prediction result at a previous time, so that the first prediction result at the previous time may be adjusted according to the second prediction result at the current time, and the adjusted result is used as the prediction result at the current time.
And step S106, generating game index prediction data of the next moment according to the first prediction result and the second prediction result.
Specifically, the first prediction result and the second prediction result are weighted according to corresponding weight values, and game index prediction data at the next moment are obtained. In practical application, according to different game scenes, the weight values of the batch processing model and the stream calculation model can be adjusted, and the method specifically comprises the following processes: responding to the weight value adjustment operation aiming at the game to be predicted, and adjusting the weight values corresponding to the first prediction result and the second prediction result respectively; and carrying out weighted calculation on the first prediction result and the second prediction result according to the adjusted weight value. For example, if the weight value of the second prediction result is increased, it indicates that the data stream at the current moment has a large influence on the game index, and if the game develops a promotion activity, the influence on the game daily income in the game index will be large, so that the user can know the change trend of the game daily income in time by adjusting the weight value of the second prediction result, and further adjust the strategy to obtain the maximum profit or cost ratio; therefore, different application scene requirements can be met by adjusting the weight values of the batch processing model and the flow calculation model.
According to the game index prediction method, the data generated by the game to be predicted can be calculated in real time by adding the flow calculation model in the prediction model, and the data of the game player can be dynamically refreshed in real time through real-time data calculation and analysis, so that the real-time prediction of the game index is realized, the timeliness and the accuracy of the prediction of the game index are improved, and the game index prediction method has a good practical value.
In practical application, the game index has the following characteristics: (1) the game index is influenced by festivals, holidays, activity promotion and the like, meanwhile, different game indexes are different in stability performance, and the daily income and the number of active users are more stable than the number of newly increased users; (2) timeliness, the life cycle of the game mainly includes stages of rising period, losing period, stable period and death period, etc., and the game indexes in different stages have timeliness, so the flow calculation model of the prediction model needs to be updated according to the actual stage.
Further, the method further comprises: and updating the flow calculation model according to the error between the second prediction result and the actual data of the game index. Specifically, the error between the second prediction result and the game index actual data is calculated, whether the error is within a preset error threshold value is judged, if yes, the streaming calculation model does not need to be updated, and if not, the streaming calculation model is updated and iterated until the error between the second prediction result output by the streaming calculation model and the game index actual data is within the preset error threshold value, so that the problem that the prediction result output by the prediction model is not ideal due to failure of the streaming calculation model is solved.
Specifically, an embodiment of the present invention further provides an updating method of a flow calculation model, as shown in fig. 2, the method includes the following steps:
step S202, obtaining second prediction results and game index actual data which respectively correspond to a specified number of moments before the current moment;
in practical application, a specified number of values can be preset, for example, the value is set to 5 times or 1 time, and the specified number of second prediction results and game index actual data before the current time are directly obtained according to the values; the number of the second prediction results may also be preset, for example, when the stream calculation model is set to update each time, N second prediction results and corresponding game index actual data are selected to perform incremental update, and specifically, the acquisition mode and the number of the second prediction results and the game index actual data may be set according to an actual situation, which is not limited herein in the embodiment of the present invention.
Step S204, calculating the difference value between the second prediction result corresponding to each moment and the actual data of the game index for each moment, and determining the increment corresponding to the current moment based on the difference value corresponding to each moment;
the difference mean value corresponding to a specified number of moments can be calculated according to the difference value corresponding to each moment, and the difference mean value is used as the increment corresponding to the current moment so as to carry out quick updating iteration on the flow calculation model; or, the difference value corresponding to each moment can be used as the increment corresponding to the current moment, so that the stream computing model can be updated and iterated according to the fine adjustment effect until the error between the second prediction result output by the updated stream computing model and the game index actual data is within the preset error threshold; and by updating the flow calculation model, dynamic correction and updating of the prediction of the future game index can be realized, so that the real-time accuracy of the game index prediction is ensured.
And step S206, updating the flow calculation model according to the increment corresponding to the current moment.
The incremental flow calculation model corresponding to the current moment is updated and iterated, so that the problem that the flow calculation model fails in different stages of the game to be predicted is solved, the prediction precision of the flow calculation model is guaranteed, in addition, the bounded data set is subjected to prediction processing through the batch processing model, the periodicity, the variation trend and the like of the prediction result can be improved, and the prediction timeliness of the game index is guaranteed; and the unbounded data set is subjected to prediction processing through the flow calculation model, so that the prediction result can be finely adjusted, such as variance and error are reduced, and the prediction accuracy of the game index is improved.
Further, an embodiment of the present invention further provides a method for training a prediction model, as shown in fig. 3, the method includes the following steps:
step S302, training an original batch processing model by applying a first data stream sample to obtain a trained batch processing model;
specifically, a first data flow sample is a bounded data set, a full amount of first data flow samples are selected to train an original batch processing model, and iteration is carried out according to preset time until a batch processing model is obtained; or calculating the error between the game index prediction result output by the original batch processing model and the game index actual data, judging whether the error is within the set index prediction error threshold value, if not, continuing iteration until the error is within the set index prediction error threshold value, and obtaining the trained batch processing model. Wherein the optimization function of the original batch model can be minimized by:
wherein M isindex,b,kRepresenting the game index prediction value output by the original batch processing model,
a mapping function representing the original batch model, T representing a time window, λ representing an optimization function of the original batch model, x0,kRepresenting the 0 th dimension at time k off-linekTrue value, x, of game index representing time k of offlinen,kRepresenting the nth dimensional feature at time k offline.
Therefore, the optimization function λ is to train the original batch processing model under the full amount of the first data stream samples, so that the difference between the predicted value of the trained batch processing model and the true value of the corresponding game index under the optimization index is the minimum.
Step S304, inputting the second data stream sample into a batch processing model to obtain a first prediction training result corresponding to the second data stream sample;
step S306, taking the first prediction training result as a true value of the second data stream sample, and training the original stream calculation model by using the true value and the unbounded data stream sample corresponding to the second data stream sample to obtain a trained stream calculation model;
specifically, after the batch processing model is trained, firstly, unbounded data stream samples corresponding to the second data stream samples are obtained, the first prediction training result is used as a reference line, namely, the reference line is used as a true value of the second data stream samples, then, the raw stream calculation model is trained according to the true value and the unbounded data stream samples, the trained stream calculation model is obtained, and an optimization function of the raw stream calculation model is optimized according to the following formula:
wherein M isindex,s,tRepresenting the game index prediction value output by the original stream calculation model,
a mapping function representing the original flow calculation model, η represents an optimization function of the original flow calculation model, xn,tRepresenting the nth dimension characteristic of t continuous time, f (t) representing the time window value of the original batch processing model corresponding to t time, Mindex,b,f(t)The first predicted training result, i.e. true value, y, of the corresponding original batch processing model at time ttRepresenting the real value of the game index when t continues.
And optimizing the optimization function of the original flow calculation model to finally obtain the trained flow calculation model. In addition, in the training process, the flow calculation model is updated in an incremental manner, which may specifically be updated according to the following formula:
wherein M isindex,s,i,jRepresenting the predicted value of the batch model output from time i to time j, l representing the mapping function of the original stream computational model, η representing the optimization function of the original stream computational model, xn,jRepresenting the nth dimension characteristic at j, f (t) representing the time window value of the original batch processing model corresponding to t, Mindex,b,f(t)The first predicted training result, i.e. true value, y, of the corresponding original batch processing model at time ttRepresenting true values of game indices at successive t, Mindex,s,i,tAnd the predicted values output by the batch processing model from the time i to the time t are shown, m represents the data quantity used for updating the flow calculation model, namely the predicted values output by m flow calculation models from the time i to the time i + m are selected, and the difference value between the m predicted values and the corresponding actual data is calculated to be used as the incremental flow calculation model for updating and iteration.
And step S308, taking the trained batch processing model and the trained flow calculation model as a prediction model of the game index.
In practical application, the trained prediction model is deployed on line, in online service, data flow is converted into a bounded data set and input into a batch processing model for prediction processing, and a first prediction result is obtained; and directly inputting the data stream into the stream calculation model for prediction processing to obtain a second prediction result, so as to obtain a prediction result of the game index according to the first prediction result and the second prediction result, storing the prediction result of the game index into a service database, and updating the iterative stream calculation model in an incremental manner after corresponding actual data of the game index is stored into service data, so that real-time prediction of the game index is realized, and the prediction accuracy is improved.
This is illustrated here for ease of understanding. As shown in the schematic diagram of fig. 4, first, a batch logic layer in an electronic device obtains a data stream corresponding to a current time of a game to be predicted, where the game related data corresponding to the current time is obtained from a log server, and game index actual data corresponding to the current time is obtained from a business database; and then, performing stream batch-to-batch calculation on the data stream to obtain a bounded data set and an unbounded data set corresponding to the data stream, and storing the bounded data set and the unbounded data set in a related database, such as a Remote Dictionary service (Redis) and an Hbase, wherein both the Redis database and the Hbase are key-value pair databases, wherein the Redis is limited by the size of a memory, but the read-write performance is very good, and the Redis is generally directly used for an online service, and the Hbase is used for storing a scene with a large data volume and is often used together with a large data component. The specific acquisition method can refer to the foregoing embodiment, and the embodiment of the present invention is not described in detail herein; it should be noted that, if only a batch processing scheme is performed, after the data stream is cut, a batch processing storage is performed through a large data-related database or a file storage system, and then a distributed computation logic is triggered at regular time to perform batch processing computation on the cut multiple data sub-streams, so as to obtain a bounded data set corresponding to the data stream.
Finally, inputting the bounded data set corresponding to the data stream into a batch processing model for prediction processing to obtain a first prediction result; and inputting the unbounded data set into the flow calculation model for prediction processing to obtain a second prediction result, so that the prediction result of the game index is obtained according to the first prediction result and the second prediction result. And updating and iterating the flow calculation model in the prediction model, and specifically setting according to the actual situation.
It should be noted that, as shown in fig. 4, in the training process of the prediction model, the trained prediction model is deployed, and in the process of online service, the prediction result of the game index of the game to be predicted is stored in the service database, and the game related data generated by the game to be predicted is stored in the log server, so that the prediction model is updated and iterated conveniently, and the prediction accuracy of the game index is ensured.
Therefore, the game index prediction method provided by the embodiment of the invention facilitates the user to select a data processing mode of batch processing or stream calculation or both, so as to generate data sets with different modes to train corresponding models and online services, and in practical application, the game index is predicted through the prediction model formed by the batch processing model and the stream calculation model, so that the method is compatible with the prediction method only performing batch processing in the existing method, and the prediction method of stream calculation can be added on the basis, thereby realizing real-time calculation of data stream and real-time reasoning of models and improving the accuracy of game indexes in prediction; meanwhile, on the level of a prediction model, latest data can be fused continuously, and the problem that the model fails in different stages of a game is solved; and the number of the prediction indexes can be transversely expanded, and the method has universality and universality for different game indexes of the game to be predicted and has better practical value.
On the basis of the method embodiment, the embodiment of the invention also provides a game index prediction device, which provides a prediction model which is trained in advance through electronic equipment, wherein the prediction model comprises a batch processing model for predicting the game index by bounded data and a flow calculation model for predicting the game index by unbounded data. As shown in fig. 5, the device includes a data stream obtaining module 51, a game index prediction module 52 and a game index prediction data generation module 53 connected in sequence, wherein the functions of the modules are as follows:
a data stream obtaining module 51, configured to obtain a data stream corresponding to a current moment of a game to be predicted; wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes;
a game index prediction module 52, configured to perform game index prediction at a next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model;
and a game index prediction data generation module 53, configured to generate game index prediction data at the next time according to the first prediction result and the second prediction result.
The game index prediction device provided by the embodiment of the invention predicts the game index of the next moment based on the data stream and the prediction model corresponding to the current moment, wherein the prediction model comprises a batch processing model and a stream calculation model.
In one possible embodiment, the game index prediction module 52 is further configured to: acquiring a bounded data set corresponding to a data stream; wherein the bounded data set carries a start time identifier and an end time identifier; inputting the bounded data set into a batch processing model of a prediction model for prediction processing; the data stream is input to a stream calculation model of the prediction model to perform prediction processing.
In another possible embodiment, the obtaining a bounded data set corresponding to a data stream includes: cutting the data stream according to a preset sliding window to obtain data sub-streams corresponding to a plurality of windows; wherein, the data sub-stream includes the start time mark and the end time mark of the window corresponding to the data sub-stream; a plurality of data sub-streams are grouped into a bounded data set.
In another possible embodiment, the game index prediction data generation module 53 is further configured to: and carrying out weighted calculation on the first prediction result and the second prediction result according to the corresponding weight values to obtain game index prediction data at the next moment.
In another possible embodiment, the performing weighted calculation on the first predicted result and the second predicted result according to corresponding weight values includes: responding to the weight value adjustment operation aiming at the game to be predicted, and adjusting the weight values corresponding to the first prediction result and the second prediction result respectively; and carrying out weighted calculation on the first prediction result and the second prediction result according to the adjusted weight value.
In another possible embodiment, the data stream obtaining module 51 is further configured to: monitoring a log file of a game to be predicted in real time to obtain game related data which influences game indexes and is generated by the game to be predicted at the current moment; acquiring game index actual data corresponding to the current moment from the service data of the game to be predicted; and forming the game related data and the game index actual data into a data stream corresponding to the current moment.
In another possible embodiment, the apparatus further comprises: and updating the flow calculation model according to the error between the second prediction result and the actual data of the game index.
In another possible embodiment, the updating the stream calculation model according to the error between the second predicted result and the actual data of the game index includes: acquiring second prediction results and game index actual data which respectively correspond to a specified number of moments before the current moment; for each moment, calculating a difference value between a second prediction result corresponding to the moment and actual data of the game index, and determining an increment corresponding to the current moment based on the difference value corresponding to each moment; and updating the flow calculation model according to the increment corresponding to the current moment.
In another possible embodiment, the training process of the prediction model includes: training an original batch processing model by applying a first data stream sample to obtain a trained batch processing model; inputting the second data stream sample into a batch processing model to obtain a first prediction training result corresponding to the second data stream sample; training a raw stream calculation model by using the truth value and the unbounded data stream sample corresponding to the second data stream sample to obtain a trained stream calculation model by using the first prediction training result as a truth value of the second data stream sample; and taking the trained batch processing model and the trained flow calculation model as a prediction model of the game index.
The game index prediction device provided by the embodiment of the invention has the same technical characteristics as the game index prediction method provided by the embodiment, so that the same technical problems can be solved, and the same technical effects can be achieved.
The embodiment of the invention also provides electronic equipment which comprises a processor and a memory, wherein the memory stores machine executable instructions capable of being executed by the processor, and the processor executes the machine executable instructions to realize the game index prediction method.
Referring to fig. 6, the electronic device includes a processor 60 and a memory 61, the memory 61 stores machine executable instructions capable of being executed by the processor 60, and the processor 60 executes the machine executable instructions to implement the game index prediction method described above.
Further, the electronic device shown in fig. 6 further includes a bus 62 and a communication interface 63, and the processor 60, the communication interface 63, and the memory 61 are connected by the bus 62.
The Memory 61 may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 63 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used. The bus 62 may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, an EISA (Enhanced Industry Standard Architecture) bus, or the like. The above-mentioned bus may be classified into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one double-headed arrow is shown in FIG. 6, but that does not indicate only one bus or one type of bus.
The processor 60 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 60. The Processor 60 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory 61, and the processor 60 reads the information in the memory 61 and, in combination with its hardware, performs the steps of the method of the previous embodiment.
The present embodiments also provide a machine-readable storage medium having stored thereon machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the game index prediction method described above.
The game index prediction method, the game index prediction device and the computer program product of the electronic device provided by the embodiment of the invention comprise a computer readable storage medium storing program codes, wherein instructions included in the program codes can be used for executing the method described in the previous method embodiment, and specific implementation can refer to the method embodiment, which is not described herein again.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In addition, in the description of the embodiments of the present invention, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (12)
1. A method of game metric prediction, wherein a pre-trained predictive model is provided by an electronic device, the predictive model including a batch model for predicting game metrics for bounded data and a flow computation model for predicting game metrics for unbounded data, the method comprising:
acquiring a data stream corresponding to the current moment of a game to be predicted; wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes;
performing game index prediction at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model;
and generating game index prediction data of the next moment according to the first prediction result and the second prediction result.
2. The method of claim 1, wherein the step of predicting the game metrics at a time next to the current time based on the data stream and the prediction model comprises:
acquiring a bounded data set corresponding to the data stream; wherein the bounded data set carries a start time identifier and an end time identifier;
inputting the bounded dataset into a batch model of the predictive model for predictive processing;
and inputting the data stream into a stream calculation model of the prediction model to perform prediction processing.
3. A game index prediction method as claimed in claim 2, wherein the step of obtaining a bounded data set corresponding to the data stream comprises:
cutting the data stream according to a preset sliding window to obtain data sub-streams corresponding to a plurality of windows; wherein, the data sub-stream contains the start time identifier and the end time identifier of the window corresponding to the data sub-stream;
a plurality of the data sub-streams are grouped into a bounded data set.
4. The game index prediction method according to claim 1, wherein the step of generating the game index prediction data of the next time based on the first prediction result and the second prediction result includes:
and performing weighted calculation on the first prediction result and the second prediction result according to corresponding weight values to obtain game index prediction data of the next moment.
5. The game index prediction method according to claim 4, wherein the step of performing weighted calculation on the first predicted result and the second predicted result according to corresponding weight values includes:
responding to a weight value adjusting operation aiming at the game to be predicted, and adjusting weight values corresponding to the first prediction result and the second prediction result respectively;
and carrying out weighted calculation on the first prediction result and the second prediction result according to the adjusted weight value.
6. The game index prediction method of any one of claims 1 to 5, wherein the step of obtaining a data stream corresponding to the current time of the game to be predicted comprises:
monitoring a log file of a game to be predicted in real time to obtain game related data which influences game indexes and is generated by the game to be predicted at the current moment;
acquiring game index actual data corresponding to the current moment from the service data of the game to be predicted;
and forming the game related data and the game index actual data into a data stream corresponding to the current moment.
7. A game index prediction method as claimed in claim 1, characterized in that the method further comprises:
and updating the flow calculation model according to the error between the second prediction result and the actual data of the game index.
8. A game index prediction method according to claim 7, wherein the step of updating the stream calculation model based on the error between the second prediction result and the game index actual data includes:
acquiring second prediction results and game index actual data which respectively correspond to a specified number of moments before the current moment;
for each moment, calculating a difference value between a second prediction result corresponding to the moment and actual data of the game index, and determining an increment corresponding to the current moment based on the difference value corresponding to each moment;
and updating the flow calculation model according to the increment corresponding to the current moment.
9. A game index prediction method as claimed in claim 1, wherein the training process of the prediction model comprises:
training an original batch processing model by applying a first data stream sample to obtain a trained batch processing model;
inputting a second data stream sample into the batch processing model to obtain a first prediction training result corresponding to the second data stream sample;
training an original stream calculation model by using the first prediction training result as a true value of the second data stream sample and using the true value and an unbounded data stream sample corresponding to the second data stream sample to obtain a trained stream calculation model;
and taking the trained batch processing model and the trained flow calculation model as a prediction model of game indexes.
10. An apparatus for predicting game metrics, the apparatus comprising, via an electronic device, a pre-trained predictive model including a batch model for predicting game metrics for bounded data and a stream computation model for predicting game metrics for unbounded data, the apparatus comprising:
the data flow acquisition module is used for acquiring a data flow corresponding to the current moment of the game to be predicted; wherein the data stream comprises: actual data of the game indexes corresponding to the current moment and game related data influencing the game indexes;
the game index prediction module is used for predicting the game index at the next moment of the current moment based on the data stream and the prediction model to obtain a first prediction result corresponding to the batch processing model and a second prediction result corresponding to the stream calculation model;
and the game index prediction data generation module is used for generating the game index prediction data of the next moment according to the first prediction result and the second prediction result.
11. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the game index prediction method of any one of claims 1 to 9 when executing the computer program.
12. A computer-readable storage medium, having stored thereon a computer program for executing the steps of the game index prediction method according to any one of claims 1 to 9 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110488896.2A CN113191810A (en) | 2021-04-30 | 2021-04-30 | Game index prediction method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110488896.2A CN113191810A (en) | 2021-04-30 | 2021-04-30 | Game index prediction method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113191810A true CN113191810A (en) | 2021-07-30 |
Family
ID=76983546
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110488896.2A Pending CN113191810A (en) | 2021-04-30 | 2021-04-30 | Game index prediction method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113191810A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113633992A (en) * | 2021-08-09 | 2021-11-12 | 网易（杭州）网络有限公司 | Game operation data prediction method and device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512762A (en) * | 2015-12-04 | 2016-04-20 | 网易（杭州）网络有限公司 | Game numerical value launching estimation method and device based on correlation analysis |
| CN107229995A (en) * | 2017-05-24 | 2017-10-03 | 腾讯科技（深圳）有限公司 | Realize method, device and computer-readable recording medium that game service amount is estimated |
| CN109815631A (en) * | 2019-02-26 | 2019-05-28 | 网易（杭州）网络有限公司 | A kind for the treatment of method and apparatus of game data |
| CN112053198A (en) * | 2020-09-21 | 2020-12-08 | 腾讯科技（深圳）有限公司 | Game data processing method, device, equipment and medium |
-
2021
- 2021-04-30 CN CN202110488896.2A patent/CN113191810A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512762A (en) * | 2015-12-04 | 2016-04-20 | 网易（杭州）网络有限公司 | Game numerical value launching estimation method and device based on correlation analysis |
| CN107229995A (en) * | 2017-05-24 | 2017-10-03 | 腾讯科技（深圳）有限公司 | Realize method, device and computer-readable recording medium that game service amount is estimated |
| CN109815631A (en) * | 2019-02-26 | 2019-05-28 | 网易（杭州）网络有限公司 | A kind for the treatment of method and apparatus of game data |
| CN112053198A (en) * | 2020-09-21 | 2020-12-08 | 腾讯科技（深圳）有限公司 | Game data processing method, device, equipment and medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113633992A (en) * | 2021-08-09 | 2021-11-12 | 网易（杭州）网络有限公司 | Game operation data prediction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112364975B (en) | Terminal running state prediction method and system based on graph neural network | |
| CN111768038B (en) | Pollutant monitoring method and device, terminal equipment and storage medium | |
| CN112734128B (en) | 7-day power load peak prediction method based on optimized RBF | |
| WO2018031958A1 (en) | Aggregate features for machine learning | |
| CN113657668A (en) | Power load prediction method and system based on LSTM network | |
| CN112308288A (en) | Particle swarm optimization LSSVM-based default user probability prediction method | |
| CN111368887B (en) | Training method of thunderstorm weather prediction model and thunderstorm weather prediction method | |
| CN110570012A (en) | Storm-based power plant production equipment fault early warning method and system | |
| CN116307215A (en) | Load prediction method, device, equipment and storage medium of power system | |
| CN110633859A (en) | Hydrological sequence prediction method for two-stage decomposition integration | |
| Shi et al. | AGG: A Novel Intelligent Network Traffic Prediction Method Based on Joint Attention and GCN‐GRU | |
| CN116845874A (en) | Short-term prediction method and device for power load | |
| CN114911788B (en) | Data interpolation method and device and storage medium | |
| CN113254472A (en) | Parameter configuration method, device, equipment and readable storage medium | |
| CN113191810A (en) | Game index prediction method and device and electronic equipment | |
| CN117494888A (en) | Distributed photovoltaic power prediction method and system based on optimized cascade | |
| CN110191015B (en) | CPI index-based cloud service performance intelligent prediction method and device | |
| CN113205230A (en) | Data prediction method, device and equipment based on model set and storage medium | |
| CN111783883A (en) | Abnormal data detection method and device | |
| CN115713044B (en) | Method and device for analyzing residual life of electromechanical equipment under multi-condition switching | |
| US11726862B1 (en) | Over-the-air machine learning | |
| CN115345346A (en) | Data prediction method and device, electronic equipment and storage medium | |
| CN112231127A (en) | Electronic device and method for analyzing reliability of equipment | |
| CN113191540A (en) | Construction method and device of industrial link manufacturing resources | |
| CN113361750A (en) | Electricity sales amount prediction method based on business expansion large data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |