Social Emotion Prediction using Lexical Normalization And Machine Learning for Roman Urdu

Project Title

Social Emotion Prediction using Lexical Normalization And Machine Learning for Roman Urdu

Project Area of Specialization Computer ScienceProject Summary

Roman Urdu is the language majorly used by people of sub-continent on different social media platforms. We know that there are a lot of websites which use emotion predictors to decide whether the review will appear on the page or not. There a lot of spelling variations of a single word of Roman Urdu as per accent and pronunciation of a person. Our proposed project will predict the emotions hidden behind the sentences and words of Roman Urdu. In this project, we will normalize the words of Roman Urdu by converting the variations of a single word into a single base from which in turn helps in a more accurate processing of Roman Urdu for Emotion Predictor.

Project Objectives

Industry Objectives

In industry, there is always a need to analyze the responses of the people over a certain product. Whether people are liking or disliking it or what changes should be made? Did the product is a success or not? The feedback given by the people plays a vital role but there is a need to normalize the given text in order to have the desired accuracy in any Roman Urdu based application like emotion detection.

Research Objectives

In chatbots and in text classification a normalizer will serve as a base tool which will be helpful while detecting emotions hidden behind the text.

Academic Objectives
It will be helpful for emotion detection while developing Q/A system in Roman Urdu for administrative purposes in institutes.

Project Implementation Method

Machine Learning Methodology

1. Data Collection

2. Data Preparation

3. Data Preprocessing

• Tokenization
• Stop Words removal
• Normalizer

4. Machine Learning Classifier

• Parameter Optimization

5. Train Model on Training Data (70%)

6. Evaluate Model on Testing Data (30%)

Benefits of the Project

Direct Customers/Beneficiaries of the Project

This will provide a platform for the users of website who are much comfortable to communicate in Roman Urdu and will help the emotion analyzer to predict their emotion and type of feedback through the text written by them.

Outputs Expected from the Project

With this Roman Urdu Normalizer, we will be able to gain desired accuracy as it will convert all the words in their base form and hence it will act as a base tool for our proposed project which is Emotion Predictor.

Technical Details of Final Deliverable

With this Roman Urdu Normalizer, we will be able to gain desired accuracy as it will convert all the words in their base form and hence it will act as a base tool for our proposed project which is Emotion Predictor.
 

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	65000
resources for data set	Equipment	5	13000	65000