https://github.com/fork123aniket/abstractive-text-summarization
Implementation of an Attention-based LSTM Encoder-Decoder Approach for Abstractive Text Summarization
https://github.com/fork123aniket/abstractive-text-summarization
abstractive-text-summarization encoder-decoder encoder-decoder-attention encoder-decoder-model keras natural-language-processing nltk nltk-python stacked-lstm tensorflow text-summarization
Last synced: 2 months ago
JSON representation
Implementation of an Attention-based LSTM Encoder-Decoder Approach for Abstractive Text Summarization
- Host: GitHub
- URL: https://github.com/fork123aniket/abstractive-text-summarization
- Owner: fork123aniket
- License: mit
- Created: 2022-11-26T12:37:16.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2022-12-06T07:16:53.000Z (over 2 years ago)
- Last Synced: 2025-01-16T06:25:59.208Z (4 months ago)
- Topics: abstractive-text-summarization, encoder-decoder, encoder-decoder-attention, encoder-decoder-model, keras, natural-language-processing, nltk, nltk-python, stacked-lstm, tensorflow, text-summarization
- Language: Python
- Homepage:
- Size: 12.7 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Abstractive Text Summarization
This repository aims at providing ***Stacked LSTM-based Encoder-Decoder*** implementation for ***Abstactive Text Summarization*** task, which generates new relevant sentence as summary for each input (original text in the form of longer paragraphs).
This task has many applications:-
- Entity timelines - Given an entity and a time period, provide a summary of the most memorable events involving this entity.
- Sentence Compression - Given a sentence, generate a shorter one while preserving the essential contextual information.
- Summarization of User Generated Content - Identify main aspects and summarize for machine or user consumption.
## Requirements
- `Tensorflow`
- `keras`
- `NLTK`
- `NumPy`
- `Pandas`
- `Matplotlib`
## Usage
### Data
- The developed model is trained on the ***Amazon Fine Food reviews*** dataset, which can be downloaded from [***here***](https://www.kaggle.com/snap/amazon-fine-food-reviews).
- The data spans a period of more than 10 years, including all ~500,000 reviews up to October 2012.
- These reviews include product and user information, ratings, plain text review, and summary.
- It also includes reviews from all other Amazon categories.
### Training and Testing
To train the model and generate summary for each new test paragraph, run `Abstractive_Text_Summarization.py`.
## Results
Following are a few summaries generated by the model:-
```
Review: aware decaf coffee although showed search decaf cups intended purchase gift kept recipient drink caffeine favorite means
Original summary: not decaf
Predicted summary: decaf decaf
```
```
Review: first time using fondarific fondant general one really easy use baby shower cake worked indicated also colored made two tier cake final product looked great greasy
Original summary: easy to use
Predicted summary: great product
```
```
Review: size quite good dog training smell strong cannot put open bag must seal everytime gave treat otherwise dog stand trying fetch believe taste great puppy purchase sure
Original summary: strong smell and my puppy loves it
Predicted summary: my dog loves these
```
```
Review: daughter drinking since months old months old still loves snack time healthy delicious great addition menu
Original summary: great snack
Predicted summary: great snack
```
```
Review: love stuff great store bought homemade baked goods kicking things professional level works colored dark light frosting also used dusting powdered sugar pretty fine texture
Original summary: fun like dust
Predicted summary: perfect
```