🚀 Excited to connect with people interested in AI · Data Science!
I’m an MTech student in AI & DS, sharing my daily learnings + side-projects here.
Looking to collaborate, learn and grow.
🔍 Open to internship / part-time roles
📬 DMs open | Let’s build together!
#connection
Hyperparameter tuning matters because:
Too deep → Overfitting 📉
Too shallow → Underfitting 📈
Good tuning helps Random Forest balance bias vs variance and improve real-world performance.
#DataScience#MachineLearning#AI#Hyperparameters#Tuning
Learning Random Forest 🌲
One of the most powerful and widely used ML algorithms for classification and prediction tasks.
Here’s a quick breakdown of how it works and the key hyperparameters that affect performance 👇
#MachineLearning#AI
Important hyperparameters ⚙️
🔹n_estimators → Number of trees
🔹max_depth → Tree depth
🔹min_samples_split → Min samples to split
🔹min_samples_leaf → Min samples in leaf
🔹max_features → Features considered per split
Good tuning helps balance underfitting vs overfitting.
🎯Completed Credit Default Risk Prediction project with SHAP
Now model predicts default risk and states which features influenced predictions most.
🔹Training, Evaluation, Deployment 🚀
🔹SHAP Explainability ✅
#MachineLearning#DataScience#AI#buildinpublic
Built Credit Default Risk Prediction project to improve my ML understanding.
Deployed on Render 🚀
🔗 https://t.co/j9SpM1ML8m
[It takes 60 secs to work after 1st API hit]
Next step: SHAP-based explainability for better interpretation 🚀
#buildinpublic#AI#ML#RandomForest
Deployed my Credit Default Risk Prediction project on Render 🚀
🔗 https://t.co/j9SpM1ML8m
[It takes 60 secs to work after 1st API hit]
Next step: adding SHAP-based explainability for better model transparency and prediction interpretation 🚀
#MachineLearning#AI#BuildInPublic
Built this Credit Default Risk Prediction (Germain Credit Dataset) project to improve my ML understanding 🚀
🔹Writing code myself
🔹Understanding the ML pipeline deeply
🔹Testing multiple models
Sometimes simple projects teach the most.
#MachineLearning#DataScience#AI
Project Update 📈
Tested:
🔹Logistic Regression
🔹Random Forest
🔹XGBoost
Random Forest currently performs best with a strong precision-recall balance and minimal overfitting for credit risk prediction 🚀
#MachineLearning#DataScience#AI
Building a Credit Default Risk Prediction project 🚀
Not a groundbreaking idea.
Goals are simple:
🔹Write more code myself
🔹Use ChatGPT less
🔹Understand ML deeply
🔹Deploy an end-to-end project
🔹Improve my resume
Basic projects teach a lot 📈
#MachineLearning#DataScience#AI
Building a Credit Default Risk Prediction project 🚀
Not a groundbreaking idea.
Goals are simple:
🔹Write more code myself
🔹Use ChatGPT less
🔹Understand ML deeply
🔹Deploy an end-to-end project
🔹Improve my resume
Basic projects teach a lot 📈
#MachineLearning#DataScience#AI
@ValoisVelvia@psomkar1 I would say, code from ChatGPT requires more prompts for perfection as compared to Claude. But yes, Claude's free usage is limited.