Home▸AI & Machine Learning▸Classifying Medical Transcriptions by Specialty with TF-IDF and Logistic Regression

🩺 Classifying Medical Transcriptions by Specialty with TF-IDF and Logistic Regression

A 3D TF-IDF document-vector space where clinical transcription notes cluster by medical specialty and a live decision boundary shows how TF-IDF plus Logistic Regression separates them.

AI & Machine Learning3DAdvanced60 FPS

A 3D document-vector space where clinical transcription notes cluster by medical specialty, with a live decision surface showing how TF-IDF weighting and a Logistic Regression classifier carve the space into specialty regions.

🔬 What It Demonstrates

Each floating point is a clinical note projected from high-dimensional TF-IDF space; the coloured floor is the classifier's decision surface. Tighter clusters and sharper boundaries mean more confident, accurate classification.

🎮 How to Use

Switch between TF-IDF and raw term-frequency features, toggle Logistic Regression vs Naive Bayes, and drag the regularization slider. Drop a new unlabelled note with "Classify new note" and watch it settle into a predicted specialty.

💡 Did You Know?

On real transcription corpora, TF-IDF + Logistic Regression is a strong, fast baseline — but it still confuses clinically adjacent specialties that share heavy vocabulary overlap, like cardiology and cardiovascular/pulmonary notes.

⚙ Under the hood

A 3D TF-IDF document-vector space where clinical transcription notes cluster by medical specialty and a live decision boundary shows how TF-IDF plus Logistic Regression separates them.

machine learningtext analysistfidflogistic regressionmedical dataclassificationThree.js

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install