CN104899413A - Team competition result prediction method based on maximum entropy model - Google Patents
Team competition result prediction method based on maximum entropy model Download PDFInfo
- Publication number
- CN104899413A CN104899413A CN201510174490.1A CN201510174490A CN104899413A CN 104899413 A CN104899413 A CN 104899413A CN 201510174490 A CN201510174490 A CN 201510174490A CN 104899413 A CN104899413 A CN 104899413A
- Authority
- CN
- China
- Prior art keywords
- model
- group
- maximum entropy
- competition
- match
- 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
Abstract
The invention relates to the field of computer data mining, in particular to a team competition result prediction method based on a maximum entropy model. The method comprises the following concrete steps of: 1, obtaining feature data of two participating parts of each team in each competition in historical competition seasons; 2, constructing the maximum entropy model according to a maximum entropy principle, and substituting a training data set T into the maximum entropy model to obtain an optimum model; and 3, obtaining an expected substituting optimum model P<*>(Y|X) of each feature datum of the two participating parts in the recent n competitions, and calculating the distribution probability P<*>(Y|X) of the finial competition result of the home field competition participating team. The technical scheme of the invention provides scientific competition result prediction for people concerning the competition.
Description
Technical field
The present invention relates to computer data excavation applications, particularly relate to a kind of group event prediction of result method based on maximum entropy model.
Background technology
Along with the development of internet and people are for the raising of the attention rate of various sports race, sports race as group's physical culture such as European football five major league, NFL, U.S.'s professional basketball league matches has not only received the concern of local resident, and these races are the universal whole world.Each large-scale betting office, large-scale website carry out interpretation and application to the result of race one after another, and the analysis that match follower provides according to the subjective preferences of oneself and website is predicted the victory or defeat of bout.But owing to affecting the many factors of group's strength, and each factor effect is not quite similar again, and the accuracy therefore predicted is very low.Be in today of large data age, we more should carry out the rational derivation of science according to existing history competition data, and can not only lean on our emotion and a small amount of data to predict the victory or defeat of bout.
" entropy " is the measure of quantity of information, and it represents that the message that a certain event occurs is more, and the possibility that event occurs is less, is exactly mathematically that probability is less.Principle of maximum entropy is pointed out, when we need to predict the probability distribution of a random occurrence, our prediction should meet all known condition, and does not do any subjectivity hypothesis to the situation of the unknown.In this case, probability distribution is the most even, and the least risk of prediction, because at this moment the information entropy of probability distribution is maximum, so people claim this model " maximum entropy model ".We know, determine that the factor of match net result may have dozens or even hundreds of kind, we carry out statistical study than the race of being better than as known data in sports, then by utilizing maximum entropy method to find an optimization model meeting thousands of kinds of different conditions simultaneously, thus match net result is predicted.
Summary of the invention
Originally Given this, main order of the present invention there are provided a kind of group event prediction of result method based on maximum entropy model.
In order to obtain above-mentioned object, the step that technical scheme of the present invention realizes:
S1 obtains the characteristic of every game competition both sides in each group history racing season.
S2, according to principle of maximum entropy, constructs maximum entropy model.By training dataset T, substitute in maximum entropy model, obtain best model.
The expectation that S3 obtains each characteristic of the competition nearest n field of both sides substitutes into optimization model
, calculate the distribution probability of the final result of the match of home court competition group
.
Further, as a kind of preferred version, the characteristic described in step S1 comprises two kinds:
1, affect the data of internal factor and the data of external factor of match net result, internal factor refers to the factor of group's therein, such as, and the repertoire of group chief coach.External factor refers to the factor of non-group therein.2, the data of the final result of the match of home court group, overall as training dataset T using characteristic.
Further, as a kind of preferred version, the structure maximum entropy model described in step S2, and solve the concrete grammar obtaining best model:
By the Liang Ge group of every history match competition, the final result of the match of home court group as the respective inherent data of random occurrence Y, Liang Ge group and external data characteristics as random occurrence X, structural environment probability distribution
.Conditional probability
entropy be: H (p)=
; According to principle of maximum entropy, training dataset T is substituted into model, obtain optimum model
.
A kind of group event prediction of result method based on maximum entropy model provided by the present invention has advantage:
The present invention mainly utilizes the historical data of match, and being obtained by principle of maximum entropy affects the internal factor of group event result and the proportion of external factor.Set up the optimal conditions probability model meeting group event, the competitive state that the group both sides that then comprehensively compete are nearest, nearest characteristic is updated in the optimization model utilizing maximum entropy model to solve, calculates the probability that competition group wins.To race follower science match prediction.
Accompanying drawing explanation
Fig. 1 is process flow diagram of the present invention;
101 S1 obtain the characteristic of every game competition both sides in each group history racing season.
102 S2, according to principle of maximum entropy, construct maximum entropy model.By training dataset T, substitute in maximum entropy model, obtain best model.
The expectation that 103 S3 obtain each characteristic of the competition nearest n field of both sides substitutes into optimization model
, calculate the distribution probability of the final result of the match of home court competition group
.
Embodiment
In order to enable the feature and advantage of above-mentioned purpose more become apparent, be described in more detail to invention below in conjunction with accompanying drawing and specific embodiments.
As shown in Fig. 1, a kind of result of the match Forecasting Methodology based on maximum entropy model provided by the present invention, is characterized in that following steps:
S1 obtains the characteristic of every game competition both sides in each group history racing season.
S2, according to principle of maximum entropy, constructs maximum entropy model.By training dataset T, substitute in maximum entropy model, obtain best model.
The expectation that S3 obtains each characteristic of the competition nearest n field of both sides substitutes into optimization model
, calculate the distribution probability of the final result of the match of home court competition group
.
Team competition chooses U.S.'s men's basketball professional tournament (NBA) in step sl, and the history racing season is wherein the competition data of 30 the whole conventional competitions of team obtaining the 2000-2014 racing season on NBA official website.Characteristic wherein: comprise every game competition team A, B internal factor two bulb separation separately, three-pointer, secondary attack, backboard, block a shot, grab, sportsman is sick and wounded, chief coach, external factor: the number of days of having a rest between team two match, weather conditions, and the victory or defeat of final home court.
In step s 2 X=(A two bulb separation, A three-pointer, A secondary attack, A backboard, A block, A grabs, A sportsman is sick and wounded, chief coach A, the number of days of having a rest between A team two match, A weather conditions, B two bulb separation, B three-pointer, B secondary attack, B backboard, B block, B grabs, B sportsman is sick and wounded, chief coach B, the number of days of having a rest between B team two match, B weather conditions), Y=(victory, bear); Composing training collection T=
.Then conditional probability distribution
conditional entropy be: H (p)=
; Training set is substituted into model, draws optimum model through training
.
Extract two bulb separations of nearest 6 matches of competition both sides A, B of being about to match, three-pointer, secondary attack, backboard in step s3, block a shot, grab, sportsman is sick and wounded, chief coach, the expectation of the characteristic such as number of days, weather conditions of having a rest between team two match substitute into optimum model
.Because Basketball Match only has dividing of victory or defeat, so only need to calculate the winning rate being about to home court team in match
.
Claims (3)
1., based on the group event prediction of result method of maximum entropy model, it is characterized in that, the method comprises the following steps:
S1 obtains the characteristic of every game competition both sides in each group history racing season;
S2, according to principle of maximum entropy, constructs maximum entropy model;
By training dataset T, substitute in maximum entropy model, obtain best model;
The expectation that S3 obtains each characteristic of the competition nearest n field of both sides substitutes into optimization model
, calculate the distribution probability of the final result of the match of home court competition group
.
2. method according to claim 1, characteristic described in wherein said step S1, it is characterized in that, characteristic comprises two kinds: the data of internal factor and the data of external factor that 1, affect match net result, and internal factor refers to the factor of group's therein, such as, the repertoire of group chief coach, external factor refers to the factor of non-group therein, 2, the data of the final result of the match of home court group, and overall as training dataset T using characteristic.
3. according to the method for claim 1 or 2, structure maximum entropy model described in wherein said step S2 also solves acquisition best model, it is characterized in that, by the Liang Ge group of every history match competition, the final result of the match of home court group is as random occurrence Y, the respective inherent data of Liang Ge group and external data characteristics as random occurrence X, structural environment probability distribution
; Conditional probability
entropy be: H (p)=
; According to principle of maximum entropy, training dataset T is substituted into model, obtain optimum model
.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510174490.1A CN104899413A (en) | 2015-04-14 | 2015-04-14 | Team competition result prediction method based on maximum entropy model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510174490.1A CN104899413A (en) | 2015-04-14 | 2015-04-14 | Team competition result prediction method based on maximum entropy model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104899413A true CN104899413A (en) | 2015-09-09 |
Family
ID=54032075
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510174490.1A Pending CN104899413A (en) | 2015-04-14 | 2015-04-14 | Team competition result prediction method based on maximum entropy model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104899413A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109165253A (en) * | 2018-08-15 | 2019-01-08 | 宁夏大学 | A kind of method and apparatus of Basketball Tactical auxiliary |
| CN109529356A (en) * | 2018-11-23 | 2019-03-29 | 腾讯科技（深圳）有限公司 | Battle result determines method, apparatus and storage medium |
| CN112272581A (en) * | 2018-01-21 | 2021-01-26 | 斯塔特斯公司 | Method and system for interactive, exposable and improved game and player performance prediction in team sports |
| CN113240190A (en) * | 2021-06-02 | 2021-08-10 | 郑州大学体育学院 | Athlete pre-race state evaluation method based on multi-period evolution entropy technology |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101645064A (en) * | 2008-12-16 | 2010-02-10 | 中国科学院声学研究所 | Superficial natural spoken language understanding system and method thereof |
| CN102034030A (en) * | 2010-12-28 | 2011-04-27 | 杭州电子科技大学 | Method for cooperatively positioning dangerous odor source by multi-robot system |
| US20120290520A1 (en) * | 2011-05-11 | 2012-11-15 | Affectivon Ltd. | Affective response predictor for a stream of stimuli |
-
2015
- 2015-04-14 CN CN201510174490.1A patent/CN104899413A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101645064A (en) * | 2008-12-16 | 2010-02-10 | 中国科学院声学研究所 | Superficial natural spoken language understanding system and method thereof |
| CN102034030A (en) * | 2010-12-28 | 2011-04-27 | 杭州电子科技大学 | Method for cooperatively positioning dangerous odor source by multi-robot system |
| US20120290520A1 (en) * | 2011-05-11 | 2012-11-15 | Affectivon Ltd. | Affective response predictor for a stream of stimuli |
Non-Patent Citations (3)
| Title |
|---|
| ADAM L.BERGER ET AL: "A Maximum Entropy Approach to Natural Language Processing", 《ASSOCIATION FOR COMPUTATIONAL LINGUISTICS》 * |
| DURSUN DELEN ET AL: "A comparative analysis of data mining methods in predicting NCAA bowl outcomes", 《INTERNATIONAL JOURNAL OF FORECASTING》 * |
| 和晶: "maxent-2.5.2", 《HTTP://WWW.PUDN.COM/DOWNLOADS208/SOURCECODE/JAVA/JAVASCRIPT/DETAIL978750.HTML》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112272581A (en) * | 2018-01-21 | 2021-01-26 | 斯塔特斯公司 | Method and system for interactive, exposable and improved game and player performance prediction in team sports |
| CN109165253A (en) * | 2018-08-15 | 2019-01-08 | 宁夏大学 | A kind of method and apparatus of Basketball Tactical auxiliary |
| CN109529356A (en) * | 2018-11-23 | 2019-03-29 | 腾讯科技（深圳）有限公司 | Battle result determines method, apparatus and storage medium |
| CN109529356B (en) * | 2018-11-23 | 2022-02-18 | 腾讯科技（深圳）有限公司 | Battle result determining method, device and storage medium |
| CN113240190A (en) * | 2021-06-02 | 2021-08-10 | 郑州大学体育学院 | Athlete pre-race state evaluation method based on multi-period evolution entropy technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Robertson et al. | A method to assess the influence of individual player performance distribution on match outcome in team sports | |
| Allen et al. | Career performance trajectories of Olympic swimmers: Benchmarks for talent development | |
| Morley et al. | An investigation of home advantage and other factors affecting outcomes in English one-day cricket matches | |
| Clarke | Home advantage in the Australian football league | |
| CN104899413A (en) | Team competition result prediction method based on maximum entropy model | |
| Carling et al. | What is the extent of exposure to periods of match congestion in professional soccer players? | |
| Mukherjee | Quantifying individual performance in Cricket—A network analysis of Batsmen and Bowlers | |
| Akhtar et al. | Forecasting test cricket match outcomes in play | |
| Mukherjee | Identifying the greatest team and captain—A complex network approach to cricket matches | |
| Goumas | Modelling home advantage in sport: A new approach | |
| Kempton et al. | The expected value of possession in professional rugby league match-play | |
| WO2015076682A1 (en) | System and method for assessing or predicting a match outcome in a sporting event | |
| Dalton-Barron et al. | A league-wide investigation into variability of rugby league match running from 322 Super League games | |
| Arboix Alió et al. | Competitive balance using Accumulated Points Difference method in male and female roller hockey leagues | |
| Sullivan et al. | Factors associated with early career progression in professional Australian Football players | |
| Munir et al. | Predicting a T20 cricket match result while the match is in progress | |
| Lu et al. | Changes in biotic and abiotic drivers of seedling species composition during forest recovery following shifting cultivation on Hainan Island, China | |
| Cao et al. | Application of tactics in technical and tactical analysis of table tennis mixed doubles based on artificial intelligence graph theory model | |
| Everson et al. | Composite Poisson models for goal scoring | |
| Patil et al. | Cricket team prediction using machine learning techniques | |
| Ramchandani et al. | India's performance in the Delhi 2010 Commonwealth Games–expected or otherwise? | |
| Saikia et al. | Quantify the fielding performance in cricket via bayesian approach | |
| Tavakol et al. | Feature Extraction and Aggregation for Predicting the EURO 2016. | |
| Sun et al. | A medal share model for Olympic performance | |
| Hamid et al. | Maximising stadium attendance: A case study of Malaysian football |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150909 |
|
| WD01 | Invention patent application deemed withdrawn after publication |