1. What can machine learning do?
Search engines, spam filtering, face recognition and so on, not only for the field of artificial intelligence, biological, medical, machinery and many other fields have been applied.
2. Definition of machine learning
A computer program was said to learn from experience E with respect to some task T and some performance measure p,if its PE Rformance on T,as measured by p,improves with experience E.--tom Mitchell
3. Supervised learning (supervised learning)
For example: In a house price forecast, we can predict prices according to different models. Supervised learning is also called regression problem, it should be said that the return is a kind of supervision, meaning to predict the output of a continuous value, such as house price prediction, tumor prediction.
4. Unsupervised learning (unsupervised learning)
In the supervised learning data set, such as tumor prediction, we can know that the data set of the tumor is benign or malignant, but in unsupervised learning of the data set is not such tags, unsupervised learning can be automatically divided into two different data sets of data clusters, so called cluster algorithm. Unsupervised learning is widely used, such as search engines to summarize the same type of information into a group, computer clusters, market segmentation, social network analysis, astronomical data analysis.
5. Cocktail party Issues
A lot of people at the party, chatting, voice, so there is a possibility: at the same time two people talking, we put two microphones in the room.
Microphone 1 Distance character 1 near a point, the microphone 2 is closer to the character 2, but the microphone can be found in the overlap of two people, which is a problem in the field of computer speech recognition, the current speech recognition technology can be high precision to identify what a person said, but when the number of speakers is two or more people, Speech recognition rates can be greatly reduced, a problem known as cocktail party issues.
6. Preparation for study
Install and familiarize yourself with the octave programming environment.
Machine Learning Study notes (1)