International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 05 Issue: 06 | June-2018 www.irjet.net p-ISSN: 2395-0072 © 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 3084 Social Network Analysis using Data Segmentation and Neural Networks Ahan M R 1 , Honnesh Rohmetra 2 , Ayush Mungad 3 1 Dept of Mathematics, BITS Pilani K K Birla Goa Campus, India 2 Dept of Computer Science & Information Systems, BITS Pilani, Pilani Campus, India 3 Dept of Computer Science & Information Systems, BITS Pilani, Pilani Campus, India -------------------------------------------------------------------------------***------------------------------------------------------------------------------ Abstract - Social Network analysis is one of the most important data analysis methods, where we gather data from the social media and network of a customer. With a lot of research and using different data segmentation and machine learning techniques we gather a deep insight of the kind of profile/category the customer belong to. Furthermore, we explore the possibility of leveraging this model and applying it in various other entities by using K-Means clustering and Natural language Processing into a multi class regression model. Further, we also interpret the model by analyzing the various attributes and use mathematical methods of Graph theory and information gain, to fit the model into an Artificial Neural network(ANN) and produce results as described in the paper. KeyWords: Machine Learning, Loan Default, Data Analysis, Data Segmentation, Graph Theory, Customer Behaviour Analysis. 1. INTRODUCTION With the advancement of technology, the data and information collected by humans, is experiencing a rapid increase. The information is greatly contributing in analysis and predictions of newer results and produce accurate strata of data, to make decisions based on data analysis algorithms. Data Segmentation is a particular topic which works mainly in segmenting data and classifying them into different classes based on the clusters formed. Different Machine Learning techniques are beneficial in designing the decision boundary to cluster the data points available. Social Network is a dense network of various data entries which can be use productively to produce meaningful results, by which a sentimental analysis along with decision trees, we can come to conclusion on the kind of customer background, the person holds and helps in easy delivery of products based on the segmentation of people. Our main aim, is to segment the target region/area of customers into classes based on significant features which would positively help the company to have a better marketing impact, therefore, minimizing their budget for the same and improving the profit by inducing this model. Graph: A graph is represented as a pair of sets (V,E), where V is the set of vertices and E is set of edges connecting different vertices. There are two types of graphs, for this particular problem, we define Multi Graph, a type of directed graph as a connection from one person to other, as it is a mutual connection. 1. Data Preprocessing This is the most important step before using any dataset, as the data needs to be processed and cleaned, by machine learning tools such as Pandas. The second important measure is to check for missing values or extreme values because they reduce the accuracy of the model and predict skewed results. If all the standard deviation values are relatively small, they can be ignored with +/- 0.5% accuracy. The third step is to normalize the data; This is an important step for all features in our dataset, as the attributes are scaled to values between (0 , 1). This step is called normalization and is an important step before processing the data through an artificial neural network.