Mushroom Classification by Physical Characteristics by Technique of k-Nearest Neighbor
Narumol Chumuang, Kittisak Sukkanchana, Mahasak Ketcham, Worawut Yimyam, Jiragorn Chalermdit, Nawarat Wittayakhom, Patiyuth Pramkeaw
Abstract
This paper proposed the principles of data analysis in order to present the prototype of mushroom classification based on physical characteristics. We created a model of mushroom classification by using Machine Learning (ML) with the mushroom dataset, comprising a total of 800 samples from the physical data of 22 attributes and it divide into two class as a toxic and non-toxic. The investigators designed the experiment in which 200 samples were randomly assigned to the mushroom population, consisting of 200 equally toxic and nontoxic mushrooms. For the quality, many ML were comparison such as Naive Bayes Updateable, Naive Bayes, SGD Text, LWL and K-Nearest Neighbor (k-NN). The results showed that K-NN gave the highest classification accuracy rate of100%.