Performance and Comparative Analysis with Incremental Neurons of ANN Hidden Layer Model on Cancer Data Set

Dr.T. Pandu Ranga Vital, Dr. Gorti Satyanarayana Murty and Dr.K. Yogeswara Rao

Now a days, the convention of (DM) data mining techniques in specifically for aging diseases like cancer, has been latterly escalating. Care ought to be taken to predict the disease and recommend the approved medicine for the pertained disease. In this research work, we focused on predict and prevent the cancer disease from Andhra Pradesh, India. The whole cancer data is collected from Andhra Pradesh (A.P.) with good questionnaire from 2016 to 2018. As per statistical analysis results, we found the cause of cancers and nature of the cancers. In this process, we conduct the performance analysis of the cancer data set and predicting the cancer disease with supervised machine learning algorithms. We applied so many Machine Learning (ML) algorithms. We are also applied ANN (Artificial Neural Networks) algorithm, it is worked effectively for using cancer dataset. In this study, we observe the results of ANN algorithm performance with 5 to 10 hidden neurons (incremental neurons in Hidden layer (HL) of ANN model). As well, we studied the performances of ML algorithms like C4.5, ADT (Alternative Decision Trees) ,Naïve Bayes and Bayes Net algorithms. In comparative study, all of ML algorithms show above 96% accuracy of the cancer data set. Bayes net performed 99.6% accuracy and C4.5 performed 99.3%.Both the algorithms take 0.05 seconds time for model construction. The ANN model with 10 hidden neurons shows the peak performance compare to all supervised machine learning algorithms. It takes 0.04 seconds for the process and gives the 100% accuracy.

Volume 11 | 04-Special Issue

Pages: 2007-2023