Software engineering, AI systems and technical paper reviews.
How I built a speech emotion recognition model in Python — parsing MFCC voice features, pruning them with mutual information, and benchmarking five classifiers (logistic regression, naïve Bayes, KNN, decision tree, MLP) in scikit-learn against WEKA.
What is logistic regression? Learn how this classification algorithm uses the sigmoid function, log loss, and gradient descent to predict probabilities and class labels.
I write about the backend×AI wedge — wins and failures included. Follow along, or say hi.