Which NLP technique is used to find the root of a word based on its part of speech?

The technique used to find the root of a word based on its part of speech is lemmatization. This process involves reducing a word to its base or dictionary form, known as a lemma, and it takes into consideration the context in which the word is used, specifically its part of speech. For example, the word "better" can be reduced to its lemma "good" when it functions as an adjective. This is crucial in natural language processing tasks where understanding the meaning derived from the context is important for applications such as text analysis, sentiment analysis, and other language applications.

In contrast, stemming is a related technique that removes suffixes from words to produce a root form but does not consider the part of speech, meaning it may not always produce a valid word. Tokenization refers to the process of breaking a text into individual words or phrases and does not involve finding root forms. Encoding typically refers to converting text data into a numerical format suitable for processing but is not related to root word identification. Thus, lemmatization is distinct in its contextual awareness and accuracy in identifying root forms based on part of speech, making it the correct choice.