Integrating Perceptron with Simple Linear Regression and Random Sampling

In recent years hybrid systems have gained more popularity in soft computing and machine learning applications. Neuro Fuzzy, Genetic Fuzzy are the most widely used hybrid techniques, by merging Fuzzy, Genetic Algorithm and Neural Network together. Hybrid System is a merging of two or more different soft computing techniques to give out a new soft computing technique. Most of the hybrid systems in soft computing consist of merging Fuzzy, Genetic Algorithm and Neural Network together. The main objective of this research is to integrate perceptron rule, simple linear regression and Random Sampling to make a new hybrid system. This new hybrid technique is applied on a famous advertising dataset, consisting of three features relating to advertising on Television, Radio and Newspaper and sales as label. Hybrid model accomplished 89 percent accuracy on train data and test data.