Customer Intent Prediction using Sentiment Analysis Techniques

Analysing the voice of the customer (VoC) through the customer intent has many applications ranging from personalised marketing to behaviour study. Individuals express their feelings in a language that is frequently accompanied by ambiguity and figure of speech, making it difficult even for humans to understand. Customer feedback is crucial as part of the customer experience (CX) management in customer retention and improves the sales strategy. Modern research has been using machine learning and word embedding technique for sentiment analysis, and it is focused on the predictive model without further context. In this study, the customer feedback comes in the form of Net Promoter Score (NPS) with a text box for written feedback. We analyse the data and demonstrate a hybrid representation that has resulted in the accuracy improvement of the sentiment classification task and predicting customer intent. The datasets were first trained using Word2Vec with the previous dataset and then fit into the Random Forest classifier, tested as the best configuration to prevent overfitting. The hybrid representation is compared against the baseline sentiment polarity tool through few experiments; the results have shown that the hybrid model has improved accuracy for the sentiment classification task. Lastly, we performed customer intent prediction by using the Power BI influencer module. The outcome of the result can be used as a reference for IT management in decision making.