Problem understanding: Text Classification using Full Convolutional NN

In this toy project, we aim to build a deep learning clssifier that predicts whether a text is a positive or negetive review. In this project you will try to use NN, namely FC NNs, for the task of text classification. Here is what we did:

• We converted the text into vectors of numbers, using different ways for word repre- sentation that can used with FC NNs.

• We explained the different ways for sentence representation • Explain different ways to deal with sentences of different length (padding, truncating, etc). • Train a FC NN for text classification. We used PyTorch for training the network. • Analyze and describe the difficulties in training FC NNs on text data.

Refer to this link to get the data.

Run the project

In this project, we have two steps: training and predicting. In the predict step, you can type a sentence as a review and predict it. Let's show you how to run the project.

If you do not have venv package, please refer to this link

Create virtual environment

$ python3 -m venv ENV_NAME

Activate your environment

$ source ENV_NAME/bin/activate

Requirement installations

To run this, make sure to install all the requirements by:

$ pip install -r requirements.txt 

Training the model

$ python3 main.py 

Results Presentation

BoW train result

TF-IDF train result

BoW predict result result

TF-IDF predict result

Make prediction

$python3 predict.py

---

---

---

Contributors

Mr. Sakayo Toadoum Sari

Mss. Fadila Hamid Abdulai

Mss. Fatoumata Haidara

Bonaventure Fonteh