A Hybrid Approach for Financial Sentiment Analysis Using Artificial Intelligence and Cuckoo Search

The newly altered read-write network has resulted in a fast development of the content written by the users that results in a large quantity of un-useful data. Opinion mining and Sentiment analysis (SA) are the two mostly used techniques in the last couple of years that processed the available online unstructured data and obtained important information as well as recognize the sentiments of user. In this research work, we will train the proposed sentiment analysis system based on Artificial Neural Network (ANN) as classification algorithm for data such as positive, negative and neutral sentiments. But ANN works mainly in two phases initially train the system as per the data available, if the training data is unique then the classification accuracy will be high so we propose cuckoo search algorithm to optimize the extracted features with a novel healthy function. On the basis of objective function, cuckoo search returns appropriate feature sets according to the data type such as positive and negative sentiments. In any sentiment analysis system work, optimization technique is needed because of unstructured processed data but in proposed work we will utilize the pre-processing steps such as normalization, stop words removal technique.