Multiclass Classification of Different Glass Types using Random Forest Classifier
Nidhi Agarwal, Rakshit Srivastava, Pratishtha Srivastava, Jassi Sandhu, Prajesh Pratap Singh · 2022 6th International Conference on Intelligent Computing and Control Systems (ICICCS) · 2022
Multiclass Classification is an inherent technique for any classification model, especially pertaining to real-life scenarios. There are various scenarios in day-to-day life where we need to classify the results based on multiple output domains. In such a scenario there is a need to have an efficient model which can proceed for multiclass classification with elevated accuracies. The research work proposed in this paper proposes a multiclass classification model with highly accurate predictions results. The results are noble to the best of our knowledge as we have been able to achieve highly meticulous predicted output values using the proposed model. The work is compared with existing values especially with the Random Forest Classification model which has the maximum accuracy till now in the case of the multiclass classification model. Our results show better prediction accuracy and other statistical values as compared to the existing work done in literature till now. The comparison is shown with all the existing approaches. All the steps for data pre-processing have been taken care of very minutely. To balance the multiclass data the SMOTE oversampling technique is applied so that the prediction accuracy is quite elevated. The real-life example which we have taken is to classify the glasses in various types so that the people involved in this manufacturing can use them most appropriately for the designing purpose.