Mushroom Classification Using Data Mining Techniques

This paper focuses on the use of classification techniques for analyzing mushroom data set. Mushroom dataset is composed of records of different types of mushrooms, which are edible or nonedible. WEKA (Waikato Environment for Knowledge Analysis) is used for implementation of the classification techniques. Different classification techniques like naïve bayes, bayes net, and ZeroR are used to categorize different mushrooms and the performance of the classification techniques is evaluated using accuracy, mean absolute error, kappa statistic. After analyzing it was found that bayes net outperformed the other techniques with highest accuracy, lowest mean absolute error and naïve bayes is the second best performer. It was also found that accuracy increased with the increase in size of the training set.