From the course: Advanced Data Processing: Batch, Real-Time, and Cloud Architectures for AI
Join today to access over 25,500 courses taught by industry experts.
Model training
From the course: Advanced Data Processing: Batch, Real-Time, and Cloud Architectures for AI
Model training
- [Instructor] The next stage in ML after feature engineering is model training. Machine learning models have evolved over time from classical ML to deep learning to generative AI. These model architectures are significantly different from one another, and so are the training processes and requirements for them. The type of model being built influences the architecture and design of the ML systems. What are the key training characteristics for different kinds of ML models? Let's begin with classical machine learning. In classical ML, we typically build supervised learning models like classification models or regression models, using techniques like Naive Bayes, decision trees, or random forests. We also do unsupervised learning like clustering. To validate the models built, we use techniques like cross validation. Metrics play a key part in the training process. For classical ML, metrics like accuracy, F1-scores, RMSE, et cetera, are used. These models require relatively small data…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.