| Nmae | ID |
|---|---|
| Ahmed Soror | 40-18290 |
| Islam Mesabah | 40-19076 |
| Abdelrahman Elbeltagy | 40-17544 |
| Islam Omar | 40-6583 |
A chatbot is a program which can do real conversations by simulating human conversations. One challenge for chatbots is recognizing feelings in the conversation partner and replying accordingly. This is a significant challenge for AI systems, however, it is straightforward for humans to recognize and acknowledge others’ feelings. One of the available datasets that helps in recognizing the feelings and improves the conversations in chatbots is the Facebook Empathetic Dialogues, they provide a novel dataset of 25k conversations grounded in emotional situations. There are no freely available similar datasets for Arabic language, however, it is the official language of 24 countries dominantly lying in the middle east, and north Africa.
The project aims to create our own Arabic dataset by translating the Facebook Empathetic Dialogues data and create an Arabic chatbot.
The trained model could be found on the drive
Dataset/Data_Cleaning.ipynb: clean the provided translated datasetData_PreProcessing.ipynb: Preprocess the dataset before trainingModel_Training.ipynb: train the model on the processed dataset and the save the modelChatbot.ipynb: Main file that opens the GUI and connects to the model.
Please see the contribution policy in the README.md on the dev branch