Pinky Thacherappillil Sabu

• Health care Data Analysis for predictive modelling (NLP Focused)

(Quarterly contribution for innovative models )

• Technologies : Python, Pandas, Scikit-learn, Jupyter Notebook, OpenAI GPT,Git,Hugging Face Transformers, SQL,NLTK,Matplotlib,Plotly,SpaCy, Seaborn,AWS SageMaker
• Project Description :

Using Natural Language Processing (NLP) Analyzed breast cancer data , patient records and clinical data to get insights and predict patient outcomes

Tokenized , removed the stop words , performed lemmatization using Spacy and NLTK

for preprocessing text data

Used Open API GPT and Hugging face transformers for applying the state -of the-art language models like NER ,text classification, sentiment analysis of medical data records.

Outcomes forecasting by building predictive models and using supervised learning with Logistic regression and BERT, DistilBERT deep learning techniques

Application of feature engineering for transforming unstructured text into structured data, medical entities , mentioning of disease and text sentiments are included from the text.

Optimized the models performance and improved the prediction accuracy by cross-validation and hyperparameter tuning

Matplotlib, seaborn , plotly used for visualization of the data

AWS sage maker used for deploying the model for real time predictions and creating user friendly clinical decision report

• Fuel Efficiency Prediction Using Linear Regression (Coursework Project)
  - Developed a linear regression model to predict fuel efficiency (miles per gallon) of vehicles based on features such as engine displacement, horsepower, and weight.Achieved a model accuracy of 85% using R-squared.
  Performed data preprocessing, including handling missing values and normalizing features. Visualized actual vs predicted values to assess model performance.
  - Skills: Python (pandas, NumPy, Matplotlib),Scikit-learn for regression modeling,Data visualization