Categorization of IRIS Flower Using Different Machine Learning Strategies

Today, machine learning is important in many areas, including agriculture, computer science, healthcare, and medicine, to name a few. It is known that our planet is abundant in species and millions of biomes thrive in the Earth's ecosystem. Some of them share physical characteristics such as size, shape, and color. It is quite difficult for a general human being to determine which items are of which sort. Similarly, it is difficult to determine which flower is of which subspecies in flower species. Classification is the most essential machine learning approach for distinguishing one item from another. Setosa, Versicolor, and Virginica are the three subspecies of the Iris flower. Five approaches are presented in depth, including Decision tree, Naive Bayes, Linear Support Vector Machine (SVM), K-nearest neighbor, and Logistic regression, and the implementation is done using the sci-kit learn the tool. Furthermore, different sizes of training and testing sets are applied to and compared with each method. The IRIS dataset was analyzed using classification and regression algorithms with a 10-fold cross-validation score. The emphasis was on examining patterns utilizing the flower's sepal and petal sizes. In terms of accuracy, the findings showed that the Linear SVM classifier outperformed with 99.17% Model accuracy and 100% of Test accuracy as compared to the Decision tree, Naive Bayes, K-nearest neighbor, and Logistic regression models.