PradipKumarDas/Competitions

This repository is the home for all competitions.

Competitions

This repository is the home for the following competitions that I have been participating in. Respective Git folders can be referred for more details.

The Competitions

• Ugam Sentiment Analysis | MachineHack: Predicting overall polarity of sentiment and aspects of the online reviews. Multi-label Classification EDA Transformers Fine Tuning

• Tabular Playground Series | Kaggle: A series of intermediate-level machine learning competitions on various tabular datasets. EDA Pipeline Cross Validation LightGBM AutoML XGBoost MLP Time Series Analysis StatsModels Keras Hyperparameters Tuning

• Shell Solar Power Prediction | HackerEarth: Predicting cloud coverage as a percentage of the open sky for a fixed field of view at 4 horizon intervals of 30, 60, 90, and 120 minutes from a 6-hour window of historical data. EDA Feature Engineering RNN Keras TensorBoard

• Data-Centric AI Competition | DeepLearning.AI: Data-centric approaches as against model-centric approaches to improves model’s prediction performance. Label Correction Noise Removal Geometric Transformations

• Car Price Prediction Hackathon | MachineHack: Predicting price of a car with different features. EDA CatBoost

• Loan Amount Prediction Competition | HackerEarth: Predicting home loan amount to be sanctioned by a bank based applicant’s demographics, financial status and the property under consideration. EDA Decision Trees Random Forests XGBoost Hyperopt

• CommonLit Readability Prize Competition | Kaggle: Assesses text readability to predict right reading level of a passage to create appropriate study content for students. EDA 1D ConvNet Transformers (BERT, RoBERTa) Cross Validation Fine Tuning

• IPL 2021 T20 Score Prediction Challenge | IIT Madras: Predicting Indian Premier League (IPL) match score at the end of six overs. EDA Preprocessing Shallow Learning Deep Learning

High-level Solution Approaches

• Exploratory Data Analysis (EDA)
• Building baseline models
• Building advanced models
• Analysing model performance
• Conclusions

Primary Techniques, Tools, Technologies & Environment

• EDA
• Feature Engineering
• Fine-tuning
• Pipeline
• Cross Validation
• Data Augmentation
• Shallow Learning
  • Linear Regression
  • Tree-based Estimators
    • Decision Tree
    • Random Forests
    • LightGBM
    • CatBoost
    • XGBoost
• Time Series Analysis [StatsModels]
• Scikit Learn
• AutoML [FLAML]
• Hyperparameters Tuning [Hyperopt]
• Artificial Neural Network & Deep Learning
  • Multi-layer Perceptrons (MLP)
  • Transformers [BERT, RoBERTa, DistilBERT]
• Keras
• TensorBoard
• Jupyter, Anaconda Data Science Platform, Google Colab
• Python
• Git
• Linux