Predictive Analytics on Facebook Data

Data in every organization are increasing at a rapid speed along with the volume of it increases at a large extend through various domains such as education, social network, meteorology, government and much more. Accordingly, Big Data refers to data as traditional data, Machine-generated Sensor data and Social data which are both structured and unstructured. Apache's Hadoop has proven to provide salient properties such as scalability, ease of use, and most notably robustness to node failures. Processing K-Nearest Neighbor queries in high dimensional data has received a lot of attention by researchers in recent years, which provides a way of predicting in a better way as done in this analysis. This research analysis is to find the means to enhance the classification accuracy, utilizing logistic regression and K-Nearest Neighbour a fast process technique in a machine learning technique for classification. The work aims at enhancing the rate of accuracy, true positive, false positive, precision, recall, sensitivity, specificity rate of classification using the HADOOP platform. As a source for analysis, political values are collected from social network users. The Prediction with K-NN neighbor approach is helpful to predict the patterns using opinion.