The dotted line in the figure is the test result of the experimental group, and the
solid line part is the test result of the control group. It can be seen that the addition of
terrain does have a certain impact on the prediction accuracy of the model, especially
when the sample size of the test set increases, it has an error of about 4% on the
prediction accuracy, but compared to other classifications. In terms of the algorithm,
adding a correction factor can greatly reduce the impact of terrain on the accuracy of
the prediction results.
Acknowledge. This work is jointly supported by the National Natural Science Foundation of
China (No. 61601082, No. 61471100, No. 61701503, No. 61750110527).
References
1. Ravari YN, Bakkes S, Spronck P (2016) Starcraft winner prediction. In: Twelfth artificial
intelligence and interactive digital entertainment conference
2. Graepel T, Minka T, Herbrich RTS (2007) A Bayesian skill rating system. Adv Neural Inf
Process Syst 19:569–576
3. Minka TP (2001) A family of algorithms for approximate Bayesian inference. Massachusetts
Institute of Technology
4. Sánchez-Ruiz AA (2015) Predicting the outcome of small battles in starcraft. ICCBR
(Workshops), pp 33–42
5. Stanescu MA, Barriga N, Buro M (2015) Using Lanchester attrition laws for combat
prediction in StarCraft. In: Eleventh artificial intelligence and interactive digital entertainment conference
6. Uriarte A, Ontanón S (2015) Automatic learning of combat models for RTS games. In:
Eleventh artificial intelligence and interactive digital entertainment conference
60
65
70
75
80
85
90
95
0
200
400
600
800
Accurancy
SIze of training set
MAP with TCF
LogisƟc Regression
MAP with TCF
LogisƟc Regression
Fig. 2. Accuracy of prediction in different situations
Battle Prediction System in StarCraft Combined with Topographic …
129
solid line part is the test result of the control group. It can be seen that the addition of
terrain does have a certain impact on the prediction accuracy of the model, especially
when the sample size of the test set increases, it has an error of about 4% on the
prediction accuracy, but compared to other classifications. In terms of the algorithm,
adding a correction factor can greatly reduce the impact of terrain on the accuracy of
the prediction results.
Acknowledge. This work is jointly supported by the National Natural Science Foundation of
China (No. 61601082, No. 61471100, No. 61701503, No. 61750110527).
References
1. Ravari YN, Bakkes S, Spronck P (2016) Starcraft winner prediction. In: Twelfth artificial
intelligence and interactive digital entertainment conference
2. Graepel T, Minka T, Herbrich RTS (2007) A Bayesian skill rating system. Adv Neural Inf
Process Syst 19:569–576
3. Minka TP (2001) A family of algorithms for approximate Bayesian inference. Massachusetts
Institute of Technology
4. Sánchez-Ruiz AA (2015) Predicting the outcome of small battles in starcraft. ICCBR
(Workshops), pp 33–42
5. Stanescu MA, Barriga N, Buro M (2015) Using Lanchester attrition laws for combat
prediction in StarCraft. In: Eleventh artificial intelligence and interactive digital entertainment conference
6. Uriarte A, Ontanón S (2015) Automatic learning of combat models for RTS games. In:
Eleventh artificial intelligence and interactive digital entertainment conference
60
65
70
75
80
85
90
95
0
200
400
600
800
Accurancy
SIze of training set
MAP with TCF
LogisƟc Regression
MAP with TCF
LogisƟc Regression
Fig. 2. Accuracy of prediction in different situations
Battle Prediction System in StarCraft Combined with Topographic …
129