🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Just used fastml to compare logistic regression (glm/glmnet) and random forest (ranger/randomForest) on the Framingham dataset. Repeated CV + Bayesian tuning (20 iter) with early stopping, MICE imputation, and upsampling made model selection easy. 🚀 library(fastml)

Looking to create stunning, data-rich maps in R? The tidyterra package makes it simple to integrate spatial data with ggplot2, bringing the power of the tidyverse to geospatial analysis. With tidyterra, you can work with spatial data just like any other data set in ggplot2. ✔️

Matrix factorizations are the pinnacle results of linear algebra. From theory to applications, they are behind many theorems, algorithms, and methods. However, it is easy to get lost in the vast jungle of decompositions. This is how to make sense of them.

Let's build an MCP server (100% locally):

Use pipelines. They instantly improve your data transformation process. 10 advantages of pipelines over the code you wrote: • Cleaner • Composable • Versionable • Testable • Reproducible • Consistent • More robust • Automatable • Modular • Scalable No brainer.

Bayes' theorem is a fundamental concept in probability theory, helping us to update our predictions or beliefs based on new evidence. It shows how conditional probabilities can be used to make better decisions by considering new data in the context of prior knowledge. The

I can't call Warp a terminal anymore. Warp is now a fully-fledged AI coding platform for your command line. They just added support to MCP, so I can literally plug in any server and have Warp do whatever I need! Here is me using Stripe and running SQL queries in English

Postdoc position in my group in Tokyo! Please get in touch if you are interested. And happy to discuss projects - ranging from developing new methods to analysis of new genomes that we are now sequencing in the lab. riken.jp/en/careers/res… Lab page: speidellab.github.io

DeepMind just dropped a 106-page paper unveiling AlphaGenome. This single model could completely redefine how we discover disease-causing mutations and drug targets. This is massive. 🧵

🚫 Stop using stepwise regression for model selection. It inflates Type I error, ignores model uncertainty, and often leads to overfitting. ✅ Instead, try: • Cross-validation • Information criteria (AIC/BIC) • Regularization (LASSO/Elastic Net) • Bayesian model averaging

I decided to put together all my MCP posts in a single PDF. It covers: - The fundamentals of MCP - Explanations with visuals and code - 11 hands-on projects for AI engineers Download link in next tweet!

The best book for Deep Learning concepts 🤯 This free book teaches practical deep learning from scratch to advanced topics like Stable Diffusion and Transformers. Covers CNNs, NLP, collaborative filtering, and more with hands-on Jupyter notebooks you can run for free in Google

This GitHub repo is a gold mine for EVERY data scientist! DS Interactive Python repo has interactive dashboards to learn statistics, ML models, and other DS concepts. Topics include PCA, bagging & boosting, clustering, neural networks, etc. Fully open-source and free!

My new book, Mathematics of Machine Learning, debuted as the #1 bestseller in the Mathematical Analysis category on Amazon. Here's its origin story and why I think you should add it to your library:

MCP is on fire. AI agents can now talk to real world tools, apps and actually get stuff done. This changes everything. Here are 10 amazing examples:

Research paper → arxiv.org/pdf/2410.24210

How to compress ML models, clearly explained (with code):

Finally! A RAG over code solution that actually works (open-source). Codebases have long-range dependencies, cross-file references that independent text chunks can't capture. Graph-Code, a graph-driven RAG system, solves this.

🚀 fastml v0.6.2 is on CRAN! New features: ✅ balance_method: Handle class imbalance with upsampling/downsampling ✅ resamples: Use custom rsample splits ✅ verbose: Get detailed training logs ✅ Improved argument validation + tuning strategy handling 📈 More control, more