SAINT CLAIR

TGIF #weekend

Building a simple neural network from scratch in java which includes Feedforward, backward propagation, Loss Function, Activation Functions github.com/Irotochukwusam… #Java #coding #AI #dev

Neural Network Feedforward ( one layer ) in Python vs Java language Python: “I’ll just guess your types and do the magic ” Java : “Type. Type. Type. Now… compile. Wait… fix that. Compile again.” #CodingLife #PythonVsJava #ProgrammerHumor

Built a tiny neural net from scratch to read handwritten digits. 0 - 9 784 inputs → 2 hidden layers → 10 outputs 1. Forward pass 2. Backprop Watching it go from random noise → recognizing numbers feels like magic! #ML #DeepLearning #Python

The pace of the AI race is forcing deeper questions. With my MSc ending in 6 months, a PhD is starting to feel less optional and more inevitable.

Mathematics, physics, and statistics aren’t just subjects….they’re the keys that unlock how the universe thinks.

@TheVixhal This is absolutely amazing man, actually built same with python but used sigmoid as activation and used Xavier initialisation to handle the vanishing gradient issue

Behind the 🎬 Hand calculating the gradient descent for multinomial logistic regression problem #tech #Ai #dev #machine #

Behind the scene of my hand calculation of feedforward propagation before I convert it into neural network python code

🚀 Building intelligence from scratch. (Neural Network) I implemented a Multi-Layer Perceptron (MLP) in Python without TensorFlow or PyTorch, to truly understand how neural networks learn. #MachineLearning #DeepLearning #NeuralNetworks #Python #AI #NapierUniversity 🧠

Just implemented k-Means clustering in Python from scratch! Used it on a small customer dataset: Income vs Spending Score. 2 clusters found automatically Centroids updated iteratively Each customer assigned to the closest centroid #MachineLearning #DataScience #Python #Cluster

Just built a Weighted K-Nearest Neighbors (KNN) Regressor from scratch using NumPy! Unlike simple KNN, this version handles : ✅ Feature Scaling (Z-score) ✅ Vectorized Distance Math ✅ Custom K selection ✅ Weighted Averaging #MachineLearning #Python #NumPy #DataScience #Coding

Built K-Nearest Neighbors from first principles for Classification • Feature scaling (Very important) • Euclidean distance (calculated by hand) • Choosing K • Weighted KNN voting #MachineLearning #python #datascience #AI #coding

I just built a Random Forest Regressor from scratch in NumPy by extending the last regression tree i built earlier. ✔️ Bootstrap sampling ✔️ Random feature selection ✔️ Variance-based splits ✔️ Recursive decision trees ✔️ Ensemble averaging #MachineLearning #RandomForest #FromScratch #NumPy #DataScience #AI #LearningML

@PriyaChakr87701 but Sklearn will be faster in larger dataset due to it Cython-optimized backend

@PriyaChakr87701 On the synthetic test data, both models achieved an MSE of ~0.813 and an R2 of ~0.96. This exists because both implementations use Variance Reduction (Mean Squared Error) as the splitting criterion and follow the CART algorithm to create a piecewise constant prediction model.

Built a Decision Tree Regressor from scratch in Python ✔ Variance-based splits ✔ Recursive tree construction ✔ Mean-value leaf predictions ✔ Piecewise constant regression #MachineLearning #DataScience #Python #FromScratch

Built a Decision Tree classifier from scratch in Python ✔ Gini impurity ✔ Threshold generation ✔ Best split selection ✔ Recursive tree building ✔ Leaf prediction & inference #MachineLearning #DataScience #Python #MLFromScratch

🚀 Built a Multinomial Logistic Regression from scratch in Python to classify breast cancer! ✅ Train/Test split ✅ One-hot encoding ✅ Softmax & gradient descent Accuracy on test set: 92.9% #python #code #ArtificialInteligence

🚀 Just implemented Ordinal Logistic Regression from scratch in Python! 🐍 ✅ Supports multiple features ✅ Learns weights, bias & thresholds ✅ Predicts ordinal outcomes (1,2,3…) ✅ Fully vectorized & easy to extend Perfect for ranking problems or ratings! #MachineLearning #Python #DataScience