Published Paper Details:

Opinion mining is crucial for recognizing and analysing opinions of users in a variety of sectors. The goal of this research is to fuse multimodal data in order to get more accurate and complete opinion mining. The code created for this project provides the user with six different choices from which to select the required functionality. Text emotion analysis, in which the code determines if textual material is good or negative. The code will examine the emotions linked with the text in the dataset after the user uploads the dataset. The thesis work builds up previous studies on opinion mining, particularly in the area of sentimental analysis and emotion detection. In the existing system, the integration of textual sentiment analysis and facial emotion analysis is not provided. The existing system contains either textual sentiment analysis or facial emotion analysis. For textual emotion analysis NLP is used and for facial emotion analysis neural networks are used. The main problem addressed by this research is the limited scope of traditional opinion mining techniques, which often rely solely on textual data and fail to consider other modalities. This leads to a less comprehensive and accurate analysis of user opinions. In the field of deep learning for face recognition, the metric commonly employed to evaluate the resemblance or dissimilarity of facial embeddings is the Euclidean distance. High-dimensional vectors based on faces that represent the individual features of a face in a learned feature space are called facial embeddings. The Euclidean distance d(x,y) between two embeddings x and y is calculated as follows: d(x,y)=?i=1n(xi-yi)2. Shannon Entropy is a measure of information uncertainty or unpredictability introduced by Claude Shannon. We can employ it to measure the information content or randomness of linguistic elements such as phonemes, words and sentences in linguistics. It is defined as: H(X)=-?i=1np(xi)log2(p(xi))