E-Commerce Product Classification using Lexical Based Hybrid Feature Extraction and SVM
V.Saravanan1, Sathya Charanya.C2
1Dr.V.Saravanan, Dean – Computer Studies, Dr.SNS Rajalakshmi College of Arts and Science (an Autonomous Institution), Coimbatore.
2Sathya Charanya.C*, Assistant professor, Department of Computer science, Salem Sowdeswari College, Self-financing Courses Wing, Salem
Manuscript received on October 17, 2019. | Revised Manuscript received on 22 October, 2019. | Manuscript published on November 10, 2019. | PP: 1885-1891 | Volume-9 Issue-1, November 2019. | Retrieval Number: L36081081219/2019©BEIESP | DOI: 10.35940/ijitee.L3608.119119
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: At present, online shopping has become a growing process, in which the profit statistics are posted by familiar e-commerce corporations like Amazon, Flipkart, Snapdeal, etc. However, this kind of online shopping unkindly omits the touch and feel of the products that can be used to estimate the product quality as the main factor while buying the commodities from the shops. The estimation of product quality is more significant during the purchasing of online products. Therefore, many opinion mining and sentiment classification methods were introduced to purchase the best products through online shopping. But, these classification methods haven’t attained the effective product classification with best reviews and ratings. In this paper, we propose a hybrid feature extraction method PCA (Principle Component Analysis) and t-SNE (t-Distributed Stochastic Neighbor Embedding ) with SVM (Support Vector Machine) using lexicon-based method to classify and separate the products from the large set of different products depending on their features, best product ratings and positive reviews. In this process, the online products will be isolated and listed according to their high positive reviews. The data preprocessing is applied to the dataset to get the data accuracy before the execution of feature extraction and classification. The dimensionality reduction and best visualization of large data set are executed by applying the PCA and t-SNE method. The sentiments are also been extracted by this hybrid feature extraction method to acquire the best neighboring product ratings. The polarity of words is discovered using a lexical based approach to extract positive reviews for obtaining the best products. Finally, the SVM is exploited to the classification of products. The performance of the proposed method is estimated with precision, recall, accuracy and complexity that can provide the entire accurateness of the system.
Keywords: Product Classification, Feature Extraction, Dimensionality Reduction, PCA, t-SNE, Lexical based Approach
Scope of the Article: Classification