Introduction to deep learning - definition Introduction (I)

Beginner little rookie , I hope it's like taking notes and recording what I've learned , Also hope to help the same entry-level people , I hope the big guys can help correct it ~ Tort made delete .

Machine learning algorithms can be divided into two categories ： Supervised learning and unsupervised learning .

One 、 Supervised learning

Supervised learning is that we teach computers how to do something . More intuitively ： Give a data set （ Training set ）, The correct answer in this data set has been set in advance , The result of supervised learning is to calculate more correct prices .

There are regression problems and classification problems in supervised learning

The regression problem is based on supervised learning to predict a continuous value output （ For example, predict the house price in a certain year ）;

Classification problem is based on supervised learning to predict discrete value output （ For example, judge whether the whole picture is a cat or a dog ）.

Two 、 Unsupervised learning

The difference between unsupervised learning and supervised learning is ： Data sets have no answer .

In unsupervised learning , There is no concept of attributes or tags , That is, all data are the same , There is no difference between .

In unsupervised learning , There is only one data set , But no one taught us how to do , We also don't know what each data point represents , Let's find out some structure by ourselves .

Typical unsupervised learning algorithms include clustering algorithm and cocktail party Algorithm

clustering algorithm ： For a given data set , Unsupervised learning algorithms may determine that this data set has multiple different clusters （ That is, self classification without correct answers ）, This is the so-called clustering algorithm .

Cocktail party algorithm ：

Let's start with the background —— Cocktail party problem ： He is a problem in the field of computer speech recognition , At present, speech recognition technology can recognize a person's words with high accuracy , When there are two or more people talking , The speech recognition rate will be greatly reduced , This problem is called the cocktail party problem .

Cocktail party algorithm is produced to solve cocktail party problems , That is to distinguish what XX said when many people spoke , What did another XX say .