Invasive Weed Optimization Based Sentiment Mining of Digital Review Content

Customer Feedback of any product or service plays an important role in the marketing industry to promote business with new features. Researcher are analyzing survey data, rating data to understand customer behavior. Text review is more effective when people express sentiment in words. This paper analyzes text review content of product / service to mine the customer sentiment. Input content was preprocessed by stopword technique and transform numeric data by use of term frequency feature. Invasive Weed Optimization Algorithm (IWOA) was used by the work to improve the content quality into two class either positive or negative review. Numeric conversion increases the fitness function efficiently to categorize txt feature into respective class. Convolution Neural Network was finally trained by improved BCS feature to reduce classification time. Experiment was done on real dataset of Amazon review. Results shows that proposed model IWCNSA (Invasive Weed and Convolution Neural Sentiment Analysis) has improved evaluation parameter values of precision, recall and accuracy.