Guide language:
I have done some simple things related to machine learning before, such as OCR recognition and license plate recognition, but I have not learned machine learning in systematic language, and recently I plan to learn it systematically. This is a study note from Ng’s machine learning. Learning to address
What is Machine Learning?
The video introduces the definition of Machine Learning given by two Machine Learning experts: 1.Artthur Samuel(1959).Machine Learning: Field of study that give computers the ability to learn without being explicitly programmed. To make a computer learn a skill in a field without explicitly programming it. 2.Tom Mitchell (1998).Well-posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E. Well-conditioned learning problem: If the performance of a computer program on T (measured by P) improves as experience E increases, then the computer program can learn some task T and some performance indicator P from experience E. Ps: It feels like the two statements are saying the same thing, although they look different. The second is more specific, and my own understanding is that computers can be trained to perform tasks in certain scenarios.
Supervised Learning:
In supervised learning, we are given a data set and already know what our correct output should look like, having the idea that there is a relationship between the input an d the output. Supervised learning problems are categorized into “regression” and “classification” problems. In a regression problem, we are trying to predict results within a continuous output, meaning that we are trying to map input variables to some cont inuous function. In a classification problem, we are instead trying to predict results in a discrete output. In other words, we are trying to map input variables into discrete categories. In supervised learning, we get a data set and already know what our correct output should look like and think that there is a relationship between the inputs and the outputs. Supervised learning problems can be divided into regression and classification problems. In regression problems, we are trying to predict the outcome in a continuous output, which means we are trying to map the input variables to some continuous function. In the classification problem, we instead try to predict the results in the discrete output. In other words, we are trying to map input variables to discrete categories. Ps: In this paper, the trend of housing price prediction is used as an example, which is relatively easy to understand. It is equivalent to some things have an inherent law. For example, housing price prediction is a quadratic equation with one yuan, x is the housing area, y is the price, and then through continuous input data to find the coefficient and index of x. After completing this formula, you can use this formula to predict the housing price of other area x. It is called supervised learning, when a person gives the calculator the right answer and guides the computer toward the desired result.
Unsupervised Learning:
Unsupervised learning allows us to approach problems with little or no idea what our results should look like. We can derive structure from data where we don’t necessarily know the effect of the variables. We can derive this structure by clustering the data based on relationships among the variables in the data. With unsupervised learning there is no feedback based on the prediction results. Unsupervised learning allows us to solve problems with little or no understanding of what the outcome should be. We can derive structures from data without knowing the effects of variables. We can derive this structure by clustering the data based on the relationships between variables in the data. In the case of unsupervised learning, there was no feedback based on the predicted results. Ps: Unsupervised learning is a process in which the computer learns by itself. Human beings do not need to give correct or wrong answers. The program automatically analyzes the same and different data to classify and distinguish them.
Conclusion:
Ng’s tutorial is in English but with bilingual subtitles, which is not difficult to understand. There is also PPT that can be downloaded, which should be very friendly to beginners like me. Combined with Zhou Zhihua’s watermelon book comparison is still very helpful. I hope I can get a new understanding of artificial intelligence and write a little project after I finish this project. China and the United States are sure to compete with each other in the coming years. The United States wants China to give up high-tech industries, including ARTIFICIAL intelligence, which arouses the enthusiasm to learn artificial intelligence. Although it is not the nature of the job, it is always good to learn more.