vaderSentiment VS wasm

There's overlap, but many traditional NLP techniques are heuristics based. Here's an example: https://github.com/cjhutto/vaderSentiment

Our code uses VADER (Valence Aware Dictionary and sEntiment Reasoner). VADER is a lexicon and rule-based sentiment analysis tool that can interpret and classify emotions.

My analyses focused on whether word usage within these threads, from 2017-2021, was positive or negative. The average level of positivity vs. negativity — often referred to as the “valence” — was scored using VADER, a language processing tool designed for online settings. Valence was averaged separately for wins and losses, then averaged again to generate a team’s overall valence score; this procedure controls for a team’s loss rate, and thus low scores do not simply reflect that a team frequently loses.

I did this all as a node project but it looks like there's a python package available here - https://github.com/cjhutto/vaderSentiment

Sources: Info on VADER, VADER Dictionary (word, score, other data), VADER's special rules, original paper by the authors.

C++, PHP, Javascript and vader

vader - a sentiment analysis tool developed by researchers at Georgia Tech

It's a fairly standard piece of machine learning called sentiment analysis. Human volunteers rate a corpus of texts as either positive or negative sentiment. A machine learning algorithm is then trained to predict those sentiment scores from the text itself. Sentiment analysis is widely used in comment and review moderation. If you use Python you can play around with open-source examples such as VADER yourself. While they can't always pick up on sarcasm or irony in general they do pretty well on picking up general trends.

You could try applying VADER (designed to handle social media data esp. Twitter) to tweets containing the word "apple" vs. "banana", and compare the sentiment scores.

We’ll download a compressed file from GitHub and extract the CSV file. The CSV file contains 25,000 rows consisting of two columns: