BERT Model and Sentiment Analysis to Identify and Analyze the Key Factors That Influence Customer Relation With Brands

Abstract

This study's main goal is to create a model that can recognise and evaluate the variables affecting consumers' interactions with brands. For this, an assorted procedures plan was used. Managers and authorities connected to the internet shopping at a creamery merchandise firm talked about the subjective component. The purpose of these interview searches is to locate and better understand the detracting parts that concede the possibility of influencing brand data on public publishing terraces. Using the Max QDA operating system, the findings of these interviews made it smooth to recognize the origin and basic codes. A dataset of 50,525 consumer-produced representations created in response to 505 advertising Instagram posts by a creamery guest was utilised in the determinable part. The BERT framework and NLP (natural language processing) methods were used to analyse and identify the feelings expressed by users in their comments. Following sentiment analysis, the sentiments were grouped into separate subjects using the K-means clustering approach.