Machine Learning
1-2 What is machine learning
- Perform on task T as measured by performance P, improve with experience E.
- Types of machine learning algorithm:
- Supervised learning → \rightarrow → teaching
- Unsupervised learning → \rightarrow →learning by itself
- Reinforcement learning
- recommender systems
- Practical advice for applying learning algorithms
1-3 Supervised learning
- The “right answer” were already given to the data set in the algorithm .
- Deal with an infinite number of features.
- Support Vector Machine, a neat mathematical track that allow a computer to with an infinite number of features.
- Regression Problem(with continuous value) or Classification Problem (with discrete value )
- Regression Problem: Predict results within a continuous output, map input variables to continuous function.
- Classification Problem: Predict the results in a discrete output, map input variables to discrete categories.
1-4 Unsupervised learning
- Given the data set we’re not told what to do with it , what each data point is and what effect of the variables.
- Clustering Algorithm , cluster individuals (based on relationships among variables) into different categories or different types.
- Cocktail Party Algorithm, find structure in a chaotic environment ( e. g. separate out two audio sources that were being summed together).