Category: Machine Learning

Machinery Fault Detection with Vibration Data Anomaly Detection using Variational Auto Encoder Auto Encoder is a unsupervised Deep Learning model. Data is encoded to a lower dimensional latent space using an encoder network. The decoder decodes the latent space data trying to reconstruct the original data. With Variational Auto Encoder (VAE), the encoder encodes probability…

Stock Portfolio Balancing with Monte Carlo Simulation Portfolio balancing is a complex optimization problem. The problem can be stated as assignment of weights to different stocks in the portfolio so that a metric called Sharpe Ratio is maximized. In this post we will see how Monte Carlo simulation provides a simple solution. The approach is brute force, because with…

The Multi-Faceted Journey of Determining ML Model Success Criteria In the world of Machine Learning (ML), defining the success criteria for your models is a crucial but often complex task. Many people tend to fixate on the idea of achieving a high percentage of accuracy, like 80% or 90%, as a one-size-fits-all benchmark. However, the…