An Improved Clustering Method For Text Documents Using Neutrosophic Logic

As a technique of Information Retrieval, we can consider clustering as an unsupervised learning problem in which we provide a structure to unlabeled and unknown data.

This book discusses the applications of different soft computing techniques for the web-based systems and services. The respective chapters highlight recent developments in the field of soft computing applications, from web-based information retrieval to online marketing and online healthcare. In each chapter author endeavor to explain the basic ideas behind the proposed applications in an accessible format for readers who may not possess a background in these fields. This carefully edited book covers a wide range of new applications of soft computing techniques in Web recommender systems, Online documents classification, Online documents summarization, Online document clustering, Online market intelligence, Web usage profiling, Web data extraction, Social network extraction, Question answering systems, Online health care, Web knowledge management, Multimedia information retrieval, Navigation guides, User profiles extraction, Web-based distributed information systems, Web security applications, Internet of Things Applications and so on. The book is aimed for researchers and practitioner who are engaged in developing and applying intelligent systems principles for solving real-life problems. Further, it has been structured so that each chapter can be read independently of the others.

Images are one of the primary media for sharing information. The image segmentation is an important image processing approach, which analyzes what is inside the image. Image segmentation can be used in content-based image retrieval, image feature extraction, pattern recognition, etc. In this work, clustering based image segmentation method used and modified by introducing neutrosophic logic.

In the last decade, opinion mining has been explored by using various machine learning methods. In the literature, document-level sentiment analysis has been majorly dealt with short-sized text only. For large-sized text, document-level sentiment analysis has never been dealt. In this paper, a hybrid framework named as ‘‘Senti-NSetPSO’’ is proposed to analyse large-sized text. Senti-NSetPSO comprises of two classifiers: binary and ternary based on hybridization of particle swarm optimization (PSO) with Neutrosophic Set. This method is suitable to classify large-sized text having more than 25 kb of size. Swarm size generated from large text can give a suitable measurement for implementation of PSO convergence. The proposed approach is trained and tested for large-sized text collected from Blitzer, aclIMDb, Polarity and Subjective Dataset. The proposed method establishes a co-relation between sentiment analysis and Neutrosophic Set. On Blitzer, aclIMDb and Polarity dataset, the model acquires satisfactory accuracy by ternary classifier. The accuracy of ternary classifier of the proposed framework shows significant improvement than review paper classifier present in the literature.

We have proposed a novel framework that performs sentiment analysis on audio files by calculating their Single-Valued Neutrosophic Sets (SVNS) and clustering them into positive-neutral-negative and combines these results with those obtained by performing sentiment analysis on the text files of those audio.

In this work, a new clustering algorithm is proposed based on neutrosophic set (NS) theory. The main contribution is to use NS to handle boundary and outlier points as challenging points of clustering methods. In the first step, a new de nition of data indeterminacy (indeterminacy set) is proposed in NS domain based on density properties of data.

This article presents a New Neutrosophic C-Means (NNCMs) method for clustering. It uses the neutrosophic logic (NL), to generalize the Fuzzy C-Means (FCM) clustering system.

In this work, a new clustering algorithm is proposed based on neutrosophic set (NS) theory. The main contribution is to use NS to handle boundary and outlier points as challenging points of clustering methods.