Модификация скрипта Python для повышения точности прогнозирования за счет интеграции алгоритмов RL NN и ML ⇐ Python

[Гость]

Hi This is my machine learning course project please help me to find a solution... ABOUT THE PROJECT The project involves modifying a python script to import a trained RL Neural Network (Deep Learning) and ML algorithms from suitable libraries with the aim of improving its prediction accuracy for football match results of the home teams. Recommended libraries for the project includes: PyTorch/TensorFlow, Scikit-learn and Keras libraries The generated predictions should be 80-100% accurate Other libraries may include Chainer. (Not more than 4 scripts) The dataset is constantly being updated to ensure success. Provide a requirements.txt file to install the necessary libraries A cost-effective approach for completing the project is recommended. AVAILABLE DATA TO BE USED AS DATASET

• cleaned matches.csv
• league table.csv (Statistical data)
• league team-form.csv INPUT AND OUTPUT The inputs are ‘cleaned matches.csv’, ‘league table.csv’ and the ‘current fixtures.csv’ to be predicted The output is ‘predicted results.csv’ TARGET, CATEGORICAL AND TEMPORAL VARIABLES ‘cleaned matches.csv’
• The target variables are ‘home_score’, ‘away_score’ and ‘winner’
• The categorical variables are ‘season_id’, ‘round’, ‘home_team’, ‘away_team’, ‘home_team_abbr’ and ‘away_team_abbr’
• The temporal Variables are ‘time’ and ‘date’ STATISTICAL DATA ‘league table.csv’
• The categorical variables are ‘Season ID’ and ‘Team’
• The temporal Variable is ‘date’
• Statistics are ‘P’, ‘Home win’, ‘Away win’, ‘W’, ‘D’, ‘L’, ‘GF’, ‘GA’, ‘DIFF’, ‘PTS’ 1 SCRIPTS TO BE MODIFIED/CREATED
• main.py: This script should be capable of importing a trained RL Neural Network and ML algorithms to generate accurate predictions by ingesting ‘current fixtures.csv’ into its dictionary and saving the generated predictions into a csv file. Additional scripts may include; a. support.py: imports additional ML algorithms into the main.py script to enhance prediction accuracy. b. neural.py: imports a trained neural network into the session to boost the prediction accuracy (if necessary). ABOUT THE FOOTBALL LEAGUE The Football league involves 16 teams that participate. There are 8 matches to be predicted every round for the home team There are 30 rounds in each season. A season has 240 matches Match Duration: 3minutes, 40 sec OBJECTIVES
• The accuracy of the prediction should be evaluated when integrating the RL Neural Network and ML algorithms into the prediction script.
• Each provided current fixtures.csv will contain at least 15 rounds of football results for the ongoing season. This is to assist the model generate accurate results for the remaining 15 rounds of the season. The required number of matches should gradually decrease as the model gets better.
• The python script should be structured to load the current fixtures.csv (season match fixtures) into a session to generate predictions.
• The predicted results should be saved into a csv file bearing the season_id number in its name I.e Predicted results 2808521.csv
• The results to be predicted by the python script are; Win, Lose or Draw PREDICTION ACCURACY After modifying the prediction script, training and tuning the RL Neural Network, it’s accuracy should improve by using Reinforcement Learning or when more data is added to the dataset. QUESTIONS, SUGGESTIONS & CORRECTIONS Any questions, suggestions about the workflow, or corrections in the creation process should be addressed before commencing this project.

Dataset and script