Performance analysis of machine learning algorithms for data classification

Dry beans are a type of edible legume that is widely farmed around the world. The quality of the seedhas an impact on crop productivity. As a result, it is critical to classify the seed for large-scale manufacturingand marketing. Raisin is also one of the most common agricultural products derived from grape drying. We usedthe dry bean and raisin datasets from the UCI ML repository in this research work, which have 16 characteristics and seven classes in the dry bean dataset with 13611 observations and 7 attributes and two classes in the raisin dataset with 900 instances. The primary goal of this research work is to offer a comparative performance analysis of Machine Learning (ML) techniques: Logistic Regression (LR), Decision Tree (DT), and Artificial Neural Network (ANN). We have used the principal components analysis PCA technique to decrease the attributes of the datasets. The performance of these three classifiers‟ was compared using dataset parameters such as size, number of attributes, number of classes, class imbalance, and missing values. The accuracy, precision, recall, and F1-score of the classifiers were also compared. According to our findings, the ANN classifier surpasses all other techniques on two datasets: dry bean and raisin, with an accuracy of 89%. The results also show that the LR and ANN perform nearly identically, whereas the DT performs the lowest of all classifiers.