In the field of natural language processing (NLP), noise estimation and removal play a critical role in refining and purifying textual data. Noise, which has the potential to obscure the true meaning of text, poses challenges for NLP algorithms. The procedure of NLP pipeline begins by identifying various types of noise in NLP applications. Tokenization, a fundamental step in preprocessing, involves breaking down text into tokens. Additionally, text preprocessing encompasses techniques such as part-of-speech tagging, parsing, stemming, stop word removal, and lemmatization. Modern NLP models necessitate text data to be presented in a vector format, a transformation achieved through feature extraction. Artificial intelligence (AI) driven NLP models harness the power of AI to autonomously remove noise within text, making them indispensable for enhancing quality of text in NLP applications.