Skip to content

N
NLP-Project
You need to sign in or sign up before continuing.
Commit 7351d642 authored by Almouhannad Hafez's avatar Almouhannad Hafez
Browse files
Options

Update results

parent 2b008fe9
Hide whitespace changes
Inline Side-by-side
Showing with 24 additions and 0 deletions
README.md
View file @ 7351d642
......@@ -7,6 +7,7 @@
**[Part1. Data preprocessing](#part1-data-preprocessing)**
**[Part2. Basic Morphological analyzer](#part2-basic-morpholgical-analyzer)**
**[Part3. Lemmatization, POS Tagging, and N-Gram](#part3-lemmatization-pos-tagging-and-n-gram)**
**[Part4. Data augmentation](#part4-data-augmentation)**
**[Results](#results)**
## ***Description***
......@@ -83,7 +84,30 @@
**`3.3.N-Grams.ipynb`**
- **Applying classification task using n-gram to perform task using only TF-IDF with different grams**
## ***Part4. Data augmentation***
**Files:**
> **`4.data_augmentation.ipynb`**
- **Applying data augmentation on the original dataset, added 5 new rephrased rows for each original row using LLM `LLama3`**
## ***Results***
> ***Using augmented dataset***
| Case\\Criterion | Accuracy(Train) | Accuracy(Test) | Precision(Test-Average) | Recall(Test-Average) | F1-Score(Test-Average) | Notes |
| ----------------------- | --------------- | -------------- | ----------------------- | -------------------- | ---------------------- | ------------------------- |
| nltk stemmer | 0.9629 | 0.9524 | 0.9513 | 0.9522 | 0.9509 | alpha=0.1, 300features |
| nltk lemmatizer | 0.9832 | 0.9699 | 0.9703 | 0.9699 | 0.9696 | alpha=0.1, 700features |
| Stanza lemmatizer | | | | | | |
| SpaCy lemmatizer | 0.9776 | 0.9657 | 0.9655 | 0.9656 | 0.9652 | alpha=0.1, 550features |
| Lemma + Verbs only | 0.7106 | 0.6321 | 0.6293 | 0.6278 | 0.6214 | alpha=0.1, 400features |
| Lemma + Adjectives only | 0.7990 | 0.7357 | 0.7383 | 0.7351 | 0.7299 | alpha=0.1, 450features |
| Lemma + Nouns only | 0.9678 | 0.9419 | 0.9406 | 0.9419 | 0.9406 | alpha=0.1, 600features |
| Text + (1,2)Gram | 0.9965 | 0.9800 | 0.9801 | 0.9799 | 0.9798 | alpha=0.01, 3100features |
| Text + (1,3)Gram | 0.9960 | 0.9807 | 0.9806 | 0.9805 | 0.9803 | alpha=0.01, 6600features |
| Text + (1,4)Gram | 0.9967 | 0.9807 | 0.9802 | 0.9805 | 0.9802 | alpha=0.01, 12100features |
| Text + (2,3)Gram | 0.9951 | 0.9695 | 0.9688 | 0.9694 | 0.9688 | alpha=0.01, 9100features |
| Text + (2,4)Gram | 0.9951 | 0.9646 | 0.9634 | 0.9645 | 0.9635 | alpha=0.01, 14100features |
> ***Applied features selection and model's hyperparameters tuning***
| Case\\Criterion | Accuracy(Train) | Accuracy(Test) | Precision(Test-Average) | Recall(Test-Average) | F1-Score(Test-Average) | Notes |
......
Results.xlsx
View file @ 7351d642
No preview for this file type
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment

HIAST - GitLab