What is machine learning?
- Arthur Samuel: The field of study that gives computers the ability to learn without being explicitly programmed.
- Tom Mitchell: A computer is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E. (一臺計(jì)算機(jī)在一系列測試T中的表現(xiàn)P，因?yàn)榻?jīng)驗(yàn)E而獲得提升)
Two broad classifications
- Supervised learning
- Unsupervised learning
Supervised learning
- Every example in our dataset has a correct answer. (因?yàn)槲覀兘o定了一定的歷史數(shù)據(jù)或者樣本數(shù)據(jù))
Supervised learning的分類：
Regression problem (回歸問題)
通過回歸來預(yù)測持續(xù)的變量
Classification problem (分類問題)
通過分類來預(yù)測離散的變量
Unsupervised learning
For unsupervised learning, we may not have or have a little data sample with the label, but we still can finish approach problem by deriving the structure form data based on relationships among the variable in the data. (對于無監(jiān)督學(xué)習(xí)產(chǎn)生的結(jié)果，我們無法對其進(jìn)行反饋，因?yàn)槲覀儧]有或者只有極少的驗(yàn)證集)
- Clustering(聚類)
- Non-clustering(非聚類)
