Python Code for AI Exercise: Assigning Sentiment Assignments for unknown words with Large Language Models

#AI Excercise

Github Link:

Nidhika1121/LLM_Sentiments: LLM_Sentiments (github.com)

Now, with this, exercise, sentiment can be assigned to the slangs in languages, even to unknown words, new words occurring in these times of app specific short words, added in our vocabulary.

In last article, we saw the AI exercise what 5 class sentiment alignment means.

However, the assignment of sentiments can be performed for existing three classes of neutral, positive and negative classes as well, this shall be covered in next article.

Here is the Python code for the 5 class sentiment alignment. The ipynb file is provided here and on Github. The codes corresponds to yesterdays article.

The results of running this code to determine the sentiment alignment of word nice, is as follows:

Neutral: Maximum Orietation 0.21501957

Bad: Maximum Orietation 0.6836092

Sad: Maximum Orietation 0.41360497

Happy: Maximum Orietation 0.5067967

Good: Maximum Orietation 1.0

The Maximum is for “nice” in the centre of cluster of the word “Good”

Note the word bad is showing a positive word too this can be set to right computations and evaluations when you compute similarity model.wv.similarity(“bad”,”good”) the value is high, I shall deal with this in coming artiles. This can be handled for automation using the right Large Language Model. 

The answer here is perfect, nice going to cluster of good, as nice is not a very high positively aligned word. 

 

Codes here:

 

 

“””AI Exercise: Sentiment_Of_Unknown_Words.ipynb

“””

 

# Mount your google drive in google colab

from google.colab import drive

drive.mount(‘/content/drive’)

import gensim

# Load Google’s pre-trained Word2Vec model.

model =  gensim.models.KeyedVectors.load_word2vec_format(‘/content/drive/MyDrive/GoogleNews-vectors-negative300.bin.gz’, binary=True)

 

vocabulary = model.wv.vocab

len(vocabulary)

model

similar_words = model.most_similar(“sad”, topn=20)

for w in similar_words:

 print(w)

 

similar_words = model.most_similar(“happy”, topn=20)

for w in similar_words:

 print(w)

 

similar_words = model.most_similar(“good”, topn=20)

for w in similar_words:

 print(w)

 

similar_words = model.most_similar(“bad”, topn=20)

for w in similar_words:

 print(w)

 

similar_words = model.most_similar(“neutral”, topn=20)

for w in similar_words:

 print(w)

 

sentiment_test = “nice”

similar_words = model.most_similar(sentiment_test, topn=10)

for w in similar_words:

 print(w[0])

 

similar_words = model.most_similar(“good“, topn=20)

for w in similar_words:

 print(w[0], “,”)

 

 

max = 0

i = 0 

index = 0

 

for w in similar_words:

 similarity = model.wv.similarity(sentiment_test, w[0])

 if max <= similarity:

   max =  similarity

   index = i

 print(“similarity between: “,  w[0],  ” and “, sentiment_test, ” is “,  similarity)

 i = i + 1 

 

print(“Maximum Orietation Towards”, similar_words[index][0], ” ” , max)

 

sentiment_test = “nice”

similar_words = model.most_similar(sentiment_test, topn=10)

for w in similar_words:

 print(w[0])

 

similar_words = model.most_similar(“happy“, topn=20)

for w in similar_words:

 print(w[0], “,”)

 

max = 0

i = 0 

index = 0

 

for w in similar_words:

 similarity = model.wv.similarity(sentiment_test, w[0])

 if max <= similarity:

   max =  similarity

   index = i

 print(“similarity between: “,  w[0],  ” and “, sentiment_test, ” is “,  similarity)

 i = i + 1 

 

print(“Maximum Orietation Towards”, similar_words[index][0], ” ” , max)

 

sentiment_test = “nice”

similar_words = model.most_similar(sentiment_test, topn=10)

 

 

for w in similar_words:

 print(w[0])

 

similar_words = model.most_similar(“sad“, topn=20)

for w in similar_words:

 print(w[0], “,”)

 

 

max = 0

i = 0 

index = 0

 

for w in similar_words:

 similarity = model.wv.similarity(sentiment_test, w[0])

 if max <= similarity:

   max =  similarity

   index = i

 print(“similarity between: “,  w[0],  ” and “, sentiment_test, ” is “,  similarity)

i = i + 1

print(“Maximum Orietation Towards”, similar_words[index][0], ” ” , max)

sentiment_test = “nice”

similar_words = model.most_similar(sentiment_test, topn=10)

for w in similar_words:

  print(w[0])

similar_words = model.most_similar(“bad“, topn=20)

for w in similar_words:

  print(w[0], “,”)

max = 0

i = 0 

index = 0

for w in similar_words:

  similarity = model.wv.similarity(sentiment_test, w[0])

  if max <= similarity:

    max =  similarity

    index = i

  print(“similarity between: “,  w[0],  ” and “, sentiment_test, ” is “,  similarity)

  i = i + 1 

print(“Maximum Orietation Towards”, similar_words[index][0], ” ” , max)

sentiment_test = “nice”

similar_words = model.most_similar(sentiment_test, topn=10)

for w in similar_words:

  print(w[0])

similar_words = model.most_similar(“neutral“, topn=20)

for w in similar_words:

  print(w[0], “,”)

max = 0

i = 0 

index = 0

for w in similar_words:

  similarity = model.wv.similarity(sentiment_test, w[0])

  if max <= similarity:

    max =  similarity

    index = i

  print(“similarity between: “,  w[0],  ” and “, sentiment_test, ” is “,  similarity)

  i = i + 1 

print(“Maximum Orietation Towards”, similar_words[index][0], ” ” , max)

Complete Code on Github

Published by Nidhika

Hi, I have an excellent problems solving skills in the domain of Engineering, and Modelling Solutions for tasks. Apart from profession, I have inherent interest in writing especially about Global Issues of Concern, poems, stories, doing painting, cooking, photography, music to mention a few! And most important on my website you can find my suggestions to latest problems, views and ideas, my poems, stories, novels, my comments, my proposals for people with funding's to implement, my blogs, my interests, my personal experiences and glimpses of my research work as well.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: