https://github.com/amber-abuah/ngram-text-generation

Text generation for autocomplete using N-Grams and Maximum Likelihood Estimators.
https://github.com/amber-abuah/ngram-text-generation

mle ngram-language-model ngrams nlp nltk streamlit

Last synced: 7 months ago
JSON representation

Text generation for autocomplete using N-Grams and Maximum Likelihood Estimators.

• Host: GitHub
• URL: https://github.com/amber-abuah/ngram-text-generation
• Owner: Amber-Abuah
• Created: 2024-11-10T21:06:21.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-11-10T23:13:12.000Z (11 months ago)
• Last Synced: 2025-01-10T20:43:01.471Z (9 months ago)
• Topics: mle, ngram-language-model, ngrams, nlp, nltk, streamlit
• Language: Python
• Homepage: https://ngram-text-generation-amber-abuah.streamlit.app/
• Size: 10.7 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
## Text Autocomplete Using N-Grams

A Streamlit app capable of auto-completing text from a given context. 

The app uses three N-grams: a Bigram, Trigram and Fourgram to generate further text from a context provided by the user.

The predicted following words are generated by Maximum Likelihood Estimator (MLE) models, trained on each N-gram.  

Below is an example of an autocomplete of the sentence _'I was in awe when I noticed'_ from each N-gram:

>**Bigram:** I was in awe when I noticed her friends in last...  

>**Trigram:** I was in awe when I noticed by the window. Emma then looked up; but as long as...   

>**Fourgram:** I was in awe when I noticed her father's house, he pleased them all.

![image](https://github.com/user-attachments/assets/4fa5bb63-25c4-48c2-8413-1190af00757b)

Training data: [Emma](https://www.gutenberg.org/ebooks/19839) and [Persuasion](https://www.gutenberg.org/ebooks/105) by Jane Austen from the [Gutenburg Corpus](https://www.gutenberg.org/).  

Libraries: `NLTK`, `Streamlit`.