Feature selection, optimization and clustering strategies of text documents

Abstract

Clustering is one of the most researched areas of data mining applications in the contemporary literature. The need for efficient clustering is observed across wide sectors including consumer segmentation, categorization, shared filtering, document management, and indexing. The research of clustering task is to be performed prior to its adaptation in the text environment. Conventional approaches typically emphasized on the quantitative information where the selected features are numbers. Efforts also have been put forward for achieving efficient clustering in the context of categorical information where the selected features can assume nominal values. This manuscript presents an in-depth analysis of challenges of clustering in the text environment. Further, this paper also details prominent models proposed for clustering along with the pros and cons of each model. In addition, it also focuses on various latest developments in the clustering task in the social network and associated environments.

Analysis of spammers’ behavior on a live streaming chat

“…In this study, we used the result of the calculation of cosine similarity [17] to represent the degree of message similarity. Cosine similarity [18][19] is the traditional method used to measure the degree of similarity between two vectors, obtained from the cosine angle multiplication. The Cosine similarity [8] can be calculated using term frequency and inverse document frequency (TF-IDF) formulas.…”

Cluster Optimization for Boundary Points using Distributive Progressive Feature Selection Algorithm

“…Feature selection approach try to find a subset of the originalvariables (also called attributes or features). In this process three different strategies can be used one is filter for information gain, wrapper is used for accuracy and embedded is used to add or remove while constructing the model based on the predicted errors [11] . In some data analysis cases such as classification or regression can be done in the reduced space more exactly than the original data space.…”

“…To measure the distance between two points Euclidean distance metric is took the major role, at the same time easily measure the data by using ruler for two and three dimensional spaces also. Sometimes Euclidean will also be selected in clustering [11] .…”

Clustering using kernel entropy principal component analysis and variable kernel estimator

“…Objects are similar inside the same cluster whereas dissimilar compared to objects descending from other clusters. Clustering, as a class of unsupervised classification method, has been widely applied in different domains, machine learning, image segmentation, pattern recognition, text mining and many other domains [1][2][3]. Great number of clustering algorithms lie in literature, the famous K-mean clustering [4], hierarchical clustering [5], k-medoids [6], and mean shift [7] have been considered in various problems.…”