Aspect and Opinion Extraction from Unstructured Data using Machine Learning Techniques

Abstract

Now-a-days, online reviews in the e-commerce website are increasingly written by the

consumers of the product. These reviews have become an important source of information for the new customers to research about these products online. The curious customer research often leads to decision making towards purchasing the product based on online reviews. These reviews reflect the experience of the user with the product over a considerable amount of time. Most of these data is saves in websites unstructured data. Retrieval and extraction of the information is essential works and importance in semantic web areas. Many of these requirements will be depend on the storage efficiency and unstructured data analysis . Merrill Lynch recently estimated that more than 80% of all potentially useful business information is unstructured data. We analyze the unstructured data individually and converted it into structured data collectively. Text mining and natural language processing are two techniques with their methods for knowledge discovery form textual context in documents. In this study, text mining and natural language techniques will be illustrated. The emerging field of opinion mining was investigated by Natural Language Processing (NLP) community for nearly two decades. In this work is inclined to feature level opinion mining of online reviews, in which the main purpose is to identify and extract product features and opinions.  Method: The step-by-step feature extrac­tion approach is followed to reach the goal of extracting maximum number of product features from the product reviews. Various types of nouns are extracted in the form of product features. These are namely frequent features, relevant features, implicit features and infrequent features. Findings: The results show that the comprehensive feature extraction approach performs better than the particular way for extracting the product features in the semantic environment. Applications: This approach is used in e-commerce websites to find out what product features are of interest to the customers. This model is useful in recommending products to the customers as the search for a product in the e-commerce site takes place, the features from the product reviews are helpful with the corresponding opinion orientations. This forms the basis for suggesting similar products using the calculated sentiments in the recommendation process