kochlisGit / ProphitBet-Soccer-Bets-Predictor

ProphitBet is a Machine Learning Soccer Bet prediction application. It analyzes the form of teams, computes match statistics and predicts the outcomes of a match using Advanced Machine Learning (ML) methods. The supported algorithms in this application are Neural Networks, Random Forests & Ensembl Models.
MIT License
281 stars 104 forks source link

Parameters Tuning #46

Open keith0811 opened 8 months ago

keith0811 commented 8 months ago

What are the parameters we can tune to increase the accuracy of the network?

KnightRider37 commented 8 months ago

What I usually do is place double chance, depending on %.

keith0811 commented 7 months ago

what train parameters can we tune to increase the accuracy ?

KnightRider37 commented 7 months ago

Well to be honest I use the program as it is. Trying to figure out the last n matches issue, as I think going back 12 years is a bit much. Many club have undergone big changes, especially investments to certain clubs which were usually in the second half of the table, now are in the top 3 or top 5. Maybe go back 5 seasons, can't really tell, as I haven't tried. I find that it's very important to look at form, as I saw many times the program giving clear winner, but that team has lost last 5-7 games and is in terrible form, which a would avoid taking as a bet. Good Luck.

molashin12 commented 7 months ago

Well to be honest I use the program as it is. Trying to figure out the last n matches issue, as I think going back 12 years is a bit much. Many club have undergone big changes, especially investments to certain clubs which were usually in the second half of the table, now are in the top 3 or top 5. Maybe go back 5 seasons, can't really tell, as I haven't tried. I find that it's very important to look at form, as I saw many times the program giving clear winner, but that team has lost last 5-7 games and is in terrible form, which a would avoid taking as a bet. Good Luck.

how can i change the dataset to just go back 5 seasons
and is there anyone tried to implement for stats to the script to use it in training and thank you for your time 😊♥️

JoeyHsu9527 commented 7 months ago

好吧,老实说,我按原样使用该程序。试图弄清楚最后 n 场比赛的问题,因为我认为回到 12 年前有点多。许多俱乐部都发生了很大的变化,特别是对某些俱乐部的投资,这些俱乐部通常处于积分榜的后半部分,现在已经进入了前3或前5名。也许回到 5 个赛季,真的说不出来,因为我没有尝试过。我发现看状态非常重要,因为我多次看到该计划给出了明显的赢家,但那支球队在过去 5-7 场比赛中输掉了比赛,状态很糟糕,人们会避免将其视为赌注。 祝你好运。

我怎样才能将数据集更改为回到 5 个赛季,是否有人试图将统计数据实现到脚本中以在训练中使用它,并感谢您抽出宝贵时间 😊♥️D:\ProphitBet-Soccer-Bets-Predictor-main\database\storage\leagues,theres a .csv file in this folder,open it and change the star year 2018

molashin12 commented 7 months ago

好吧,老实说,我按原样使用该程序。试图弄清楚最后 n 场比赛的问题,因为我认为回到 12 年前有点多。许多俱乐部都发生了很大的变化,特别是对某些俱乐部的投资,这些俱乐部通常处于积分榜的后半部分,现在已经进入了前3或前5名。也许回到 5 个赛季,真的说不出来,因为我没有尝试过。我发现看状态非常重要,因为我多次看到该计划给出了明显的赢家,但那支球队在过去 5-7 场比赛中输掉了比赛,状态很糟糕,人们会避免将其视为赌注。 祝你好运。

我怎样才能将数据集更改为回到 5 个赛季,是否有人试图将统计数据实现到脚本中以在训练中使用它,并感谢您抽出宝贵时间 😊♥️D:\ProphitBet-Soccer-Bets-Predictor-main\database\storage\leagues,theres a .csv file in this folder,open it and change the star year 2018

thanks pro if u have time can i ask u i can't get a model accuracy more than 64 in premier league is this ok or I can getbetter outcome and how please i am usin N = 5 league from 2018 untill now auto training Rf and NN on 50 evaluation sample and i traied multiple values of trials and same outcome

JoeyHsu9527 commented 7 months ago

好吧,老实说,我按原样使用该程序。试图弄清楚最后 n 场比赛的问题,因为我认为回到 12 年前有点多。许多俱乐部都发生了很大的变化,特别是对某些俱乐部的投资,这些俱乐部通常处于积分榜的后半部分,现在已经进入了前3或前5名。也许回到 5 个赛季,真的说不出来,因为我没有尝试过。我发现看状态非常重要,因为我多次看到该计划给出了明显的赢家,但那支球队在过去 5-7 场比赛中输掉了比赛,状态很糟糕,人们会避免将其视为赌注。 祝你好运。

我怎样才能将数据集更改为回到 5 个赛季,是否有人试图将统计数据实现到脚本中以在训练中使用它,并感谢您抽出宝贵时间 😊♥️D:\ProphitBet-Soccer-Bets-Predictor-main\database\storage\leagues,theres a .csv file in this folder,open it and change the star year 2018

thanks pro if u have time can i ask u i can't get a model accuracy more than 64 in premier league is this ok or I can getbetter outcome and how please i am usin N = 5 league from 2018 untill now auto training Rf and NN on 50 evaluation sample and i traied multiple values of trials and same outcome