Question Type Classification using Machine Learning (SVM)

Project Title: Question Type Classification using Machine Learning (SVM)

Description:

This project focuses on building a Natural Language Processing (NLP) model to automatically classify questions into their respective types such as WHAT, WHO, WHEN, WHERE, WHY, HOW, WHICH, and OTHER.

The system uses text preprocessing techniques combined with feature extraction and a machine learning model to achieve accurate classification results.

⚙️ Key Features:

? Automatic Labeling:

Questions are labeled based on their starting word using regex patterns.

? Text Preprocessing:

Lowercasing

إزالة الرموز (punctuation)

إزالة الـ stopwords

Stemming باستخدام PorterStemmer

Lemmatization باستخدام WordNetLemmatizer

? Feature Extraction:

استخدام TF-IDF Vectorizer

دعم Unigrams & Bigrams

? Model Training:

استخدام Support Vector Machine (SVM) عبر LinearSVC

تدريب الموديل على بيانات train

? Evaluation Metrics:

Accuracy

Precision

Recall

F1 Score

Classification Report

? Visualization:

رسم Bar Chart لعرض أداء الموديل

? Model Saving:

حفظ:

الموديل

الـ Vectorizer

الـ Label Encoder

باستخدام joblib

? Output:

ملف CSV يحتوي على:

البيانات الأصلية

نوع السؤال المتوقع (Predicted Label)

? Technologies Used:

Python

Pandas

NLTK

Scikit-learn

Matplotlib

Joblib

? Project Goal:

To build an efficient and automated system capable of understanding and classifying question types, which can be useful in:

Chatbots ?

Question Answering Systems

Search Engines

? Future Improvements:

استخدام Deep Learning (زي LSTM أو BERT)

تحسين الـ preprocessing (زي التعامل مع السياق)

زيادة حجم الداتا لتحسين الأداء