Predicting Mushroom Edibility with Effective Classification and Efficient Feature Selection Techniques

Abstract

Machine learning in agriculture has added a new dimension to research field through providing actionable insights for better crop yield. In our work, we have explored various types of mushrooms and utilized efficient feature selection techniques coupled with effective machine learning based classification methods to classify the edibility of mushrooms automatically. Through experimentation with nine distinct machine learning methods and twenty selected features, the result shows that our best model (k-NN) performed significantly better with an accuracy of 99% and F-1 score of 99% in comparison to other machine learning methods. Our research work provided us with valuable actionable insights to devise further scopes in research in the field of mushroom edibility predicting through machine learning.