NLP Ssentiment analysis & Machine Learning

Methodology: Data preprocessing: This involves tokenization and removing punctuations or any symbols within the sentence. Other elements removed are the stop words and the links to the websites or other social media platforms provided by the user. Transformation: Considering that machine learning model accepts numbers only, it will be necessary to convert the sentences generated into numbers. This can be done using the TF-IDF technique. Here, tokenization is done to the sentences into Bag of Features. After this step, data preprocessing is conducted again with the concern being on the removing any punctuations available in the data sets. The sentences are checked for any idiom or phrase lexicon, in which the researcher replaces them with the score generated from lexicon. With this in place, it helps in improving the uniformity of the data analysis. Here, sentences are tokenized into Bag of Word. Finally, the sentiment score is calculated using the lexicon resources and machine learning score towards predicting the final score for the sentiment analysis of the sentences. Ressources: Sentiment Lexicons – SentiWordNet resource will be accessed, which is used in opinion mining, and AFINN wordlist for conducting sentiment analysis. Machine learning – TensorFlow, which is an open-source machine learning, and PyToch for social media deep learning of the sentiment analysis. Natural language processing (NLP) tools – this includes a wide range of tools such as the natural language toolkit (NLTK), and spaCY (to access advanced natural language processing using Python). Dataset – Twitter sentiment analysis dataset. BERT Pre-trained model to be used in the processing of different natural language tasks.