The classification of images into semantic categories is tough nowadays. This paper presents a system to classify real world scenes in four semantic groups of coast, forest, highways and street using support vector machines. Established classification approaches simplify badly on image classification tasks, when the classes are non-separable. In this paper we used Support Vector Machine for scene classification. Support Vector Machine is a supervised classification technique, has its extraction in geometric Learning Theory and have gained importance as they are strong, precise and are effective even after using a small training model. With their character Support Vector Machines are basically binary classifiers, though, they can be tailored to handle the manifold classification tasks general in scene classification. This proposed work shows that support vector machines can simplify well on hard scene classification problems. Support Vector Machines can execute well on a non-linear classification using kernel deception, completely mapping their inputs into high-dimensional feature spaces. In this paper 3 types of kernels (linear, polynomial and RBF kernels) are used with support vector machines. It is observed that Gaussian kernel outperform other types of kernels.
Social media mining is the process of representing, analyzing, and extracting actionable patterns and trends from raw social media data. Social media is favored by many users since it is available to individuals without any limitations to share their opinions, educational learning experiences and concerns via their status. Twitter API, twitter4j, is processed for searching the tweets based on the geo location. Student's posts on social network offers us a stronger concern to take decisions concerning the particular education system's learning method of the system. Evaluating knowledge in social media is sort of a difficult method. Bayes classifier are enforced on deep-mined knowledge for analysis purpose to urge the deeper understanding of the information. It uses multi label classification technique as every label falls into completely different classes. Label based measures are mostly taken to research the results and comparing them with the prevailing sentiment analysis technique.
30 April 2016 in International Journal of Security and Its Applications