Prakatheeswari Ravi

• Created a machine learning model utilizing Transfer Learning (MobileNet) to categorize the emotion in an image. Implemented the model in python and achieved an accuracy of 94%.
• Utilized FER dataset from Kaggle to train the model, comprised of 7 emotion categories. By deploying this model, it is possible to recognize human emotions in real-time.

To Build a Machine Learning Model, which will help to shortlist the potential candidates, based on certain criteria. The model is deployed in the cloud.
Tools used: Python,
Techniques Used: Logistic Regression in Machine Learning and Flask for deploying model in the cloud.

Presented paper in National Conference on Machine Learning and Artificial Intelligence conducted by Indian. IIM Bangalore and CIT Coimbatore. Performed regression analysis to infer whether the cancer rates have been influenced or not.

Performed Extensive Data cleaning and analyzed the patterns on credit score dataset using Clustering, Regression,
Classification and Visualization (Seaborn) techniques to establish three interesting findings.
Trained a Decision Tree model, Random Forest model and K-NN classification to predict the credit score and performed a
comparative study using AUC score as the metric.

- Conducted exploratory data analysis to identify patterns and trends in data
- Performed feature engineering to extract meaningful information from features
- Built machine learning models using algorithms such as linear regression, random forest, xg boost, lasso, ridge, and gradient boosting
- Experimented with ensemble models such as stacking and blending to improve performance
- Evaluated each model's performance using metrics such as mean squared error and root mean squared… Show more

- Conducted exploratory data analysis to identify patterns and trends in data
- Performed feature engineering to extract meaningful information from features
- Built machine learning models using algorithms such as linear regression, random forest, xg boost, lasso, ridge, and gradient boosting
- Experimented with ensemble models such as stacking and blending to improve performance
- Evaluated each model's performance using metrics such as mean squared error and root mean squared error
- Achieved a final rank of 243 in the competition after multiple iterations and rounds of model training and evaluation. Show less

Implemented a CNN classification model using TensorFlow and Keras to predict image classes with an accuracy of 94%.
Optimized the model's performance through data preprocessing techniques such as scaling, data splitting, and augmentation

• Developed a predictive model to find out the sales of a product at a particular store using various model and conducted a
comparative study.
• The main objective is to compare the Linear regression model, Decision Tree Model, Random Forest and Extra Tree
Regressor leveraging CV Score as the metric.
• Built an interactive dashboard in PowerBI to gain deeper insight about the data.

Implemented Sentiment Analysis on twitter data using Natural Language processing. Applied count vectorizer technique to extract features for processing the model. Built a logistic regression model to classify the sentiment of tweet.
• Retrieved live tweets using twitter API. Performed Exploratory data analysis on data to understand the features, for anomaly detection and assigned the target value by calculating the positive and negative word count.

-- Analyzed Walmart's sales data using advanced statistics for strategic insights. Investigated trends and patterns, and identified factors such as holidays and temperature that impact sales.
-- Suggested ways for Walmart to improve sales based on insights gained from data analysis. Developed accurate predictions using generalized linear models and clustering analysis.
-- Identified key factors that affect sales performance for Walmart. Used statistical methods to provide data-driven… Show more

-- Analyzed Walmart's sales data using advanced statistics for strategic insights. Investigated trends and patterns, and identified factors such as holidays and temperature that impact sales.
-- Suggested ways for Walmart to improve sales based on insights gained from data analysis. Developed accurate predictions using generalized linear models and clustering analysis.
-- Identified key factors that affect sales performance for Walmart. Used statistical methods to provide data-driven recommendations to improve Walmart's sales performance. Show less