Very insightful paper(https://t.co/WCbSibCwux) for comparison between XGBoost and Logistic Regression (LR) for binary classification tasks, particularly in imbalanced datasets like burned area detection from satellite imagery.

Some key Points Summarizing the comparison: pic.twitter.com/tFfjq8Bv9h

— Śreyāśḥ D 🚀 𝗲/𝗮𝗰𝗰 (@xegression) July 20, 2025

In 1886, Francis Galton observed that children’s heights tended to be closer to the average than their parents’.

This “regression toward the mean” led to the term we now use in statistics: regression. pic.twitter.com/ezkn6w9bDO

— Śreyāśḥ D 🚀 𝗲/𝗮𝗰𝗰 (@xegression) August 5, 2025

One of the most cited papers, which introduced random forest by Leo Breiman (though idea was there long ago).

He also has website for it and just got to know that Random Forest itself is not patented,but the name is trademarked. https://t.co/X0fXWL4uWJhttps://t.co/AM5wpPSRMC pic.twitter.com/NH5eQ5H4M9

— Śreyāśḥ D 🚀 𝗲/𝗮𝗰𝗰 (@xegression) July 16, 2025

If u are tired of ML predicting models, here are some project/paper ideas
📘 Study the following material( lecture+paper+code) on martingale posterior distributions and
✍️ Write a paper applying the framework to path-dependent European options ( e.g., lookback options)
1/3 pic.twitter.com/TXKx9DJ9bT

— Śreyāśḥ D 🚀 𝗲/𝗮𝗰𝗰 (@xegression) August 5, 2025

It always amazes me how one day George Dantzig arrived late to a statistics class, mistook two unsolved problems on the board as homework, and ended up proving them, which ultimately became his PhD thesis.
might be one of the fastest thesis completions in academic history. pic.twitter.com/SC8S0IHTK0

— Śreyāśḥ D 🚀 𝗲/𝗮𝗰𝗰 (@xegression) June 28, 2025