Real Data, Scientific Models, Real Results
Prediction Performance Keep improving
2024-03-24
Today's race results are exciting π, partly due to several days of big data testing. Monkey tested hundreds of parameter combinations. Based on the data results, the existing models were optimized. At the same time, several underperforming models were also eliminated, keeping only the best. Today's results did not disappoint Monkey. Both models, RandForest and NeuroNet, achieved a good hit rate of 47%, while the other two models also maintained a hit rate of 43%. Hopefully, the results can be maintained.
Although the results have improved, computers still lost to humans today. This time, AliceWin didn't take profits as she did last time but fully armed herself, participating in a total of ten races. Leading all the way, with a hit rate of 50% at best, she secured the top spot, which is commendable.
Member Nj313m also performed bravely this time, with results advancing significantly compared to before, achieving a hit rate of 47% and reaping profits. It can be seen that participating in the horse selection game on this site, everyone can learn from each other, which will significantly help everyone's horse selection ability.
Until next time!
About Monkey
The monkey believes that facts and science are the best ways to make decisions and predictions. That's why it created this website, to make different types of machine learning predictions and verify their accuracy. Although many people say that horse racing is unpredictable and random, I want to see what level of accuracy can be achieved by applying machine learning to this data.
Popular Posts
- Horse Matchup Chart
- Human Win over Machine This Time
- Real Data, Scientific Models, Real Results
- Machine Learning Model Won again
- Treasure Hunting in the Accuracy Board
- Human vs Machine Learning
- Limitation of Prediction Model
- Horse Racing (Random Forest)
- Horse Racing (Logistic Regression)
- Machine Learning for Horse Racing (Naive Bayes)