POLI 210 Lecture Notes - Lecture 17: Natural-Language Processing, Sentiment Analysis
21 Feb 2018
School
Department
Course
Professor
Document Summary
Want to discover the meaning from texts. Uses techniques such supervised, unsupervised, dictionary, regression. From this we can assume that the tweets from trumps twitter account that are sent in the morning are from him and the ones in the afternoon are from his campaign. We notice that donald trump tweets (android) is that his words used are more negative. The words from the campaign (iphone) are mostly neutral words and promotes his campaign. for ex: #trumppence16 #makeamaericagreatagain. Strip out un useful information and standardize texts. Get rid of structure and just use words: n-grams. Use combination of words. sequences of words present in the text. For ex: measure presence or absence of jail hilary: named entity recognition. Extracts named entity of people or companys. We want to remove these when analyzing data. Strip white-space and remove numbers and stem. All words ending in the letter y end with i for ex: baby = babi.
