Earlier, we mentioned supervised learning, which corresponds to non-supervised learning in machine learning. The problem with unsupervised learning is that in untagged data, you try to find a hidden structure. Because the examples provided to learners arenot marked, so there is no error or reward signal to evaluate the potential solution. This differs from supervised learning and intensive learning unsupervised learning. Unsupervised learning is closely related to the problem of density estimation of statistical data. Unsupervised learning, however, includes techniques for seeking, summarizing, and interpreting key features of data. Many of the methods used in unsupervised learning are based on data mining methods used to process data. Let's take a look at two pictures: and nbsp , &NB Sp we can see that unsupervised learning does not have any tags or tags or tags. So we know the data set, but we don't know what to do with it, and we don't tell what each data point is. Nothing else, just a data set. For data sets, unsupervised learning can determine that there are two different clusters of data. This is one, that is the other, the two are different. Unsupervised learning algorithms may divide these data into two different clusters. So it's called a clustering algorithm. It turns out that it can be used in many places. One example of cluster application is in Baidu News. If you have never seen it before, you can go to this URL http://news.baidu.com/to see it. Baidu News every day in, collect very many, very many network news content. It then groups the news together, forming the associated news. So Baidu News to do is to search for a lot of news events, automatically put them together. So, these news events are all on the same subject, so they show up together. , &NB Sp , &NBSp From this page, Baidu News collects a lot of news, then gathers them into different classes, for example: Real estate, Internet ... In each big class (Big label), and then together into a different small class. Let's look at another example: an example of a DNA microscopic data. , &NB Sp , &NB Sp The basic idea is to enter a set of different individuals, for each of them, you have to analyze whether they have a specific gene. Technically, you have to analyze how many specific genes have been expressed. So these colors, red, green, gray, and so on, these colors show the degree to which different individuals have a specific gene. All you can do is run a clustering algorithm that will cluster individuals into different classes or different types of groups (people) ... So this is unsupervised learning, because we didn't tell the algorithm in advance, for example, this is the first class of people, those are the second class, the third class, and so on. We're just saying that this is a bunch of data. I don't know what the data is, I don't know who it is, I don't even know what the different types of people are, and what these types are. But can you automatically find the structure in the data? It means that you want to automatically cluster those individuals into various classes, and I can't know in advance which ones. Because we did not give the algorithm the correct answer to respond to data in the dataset, this is unsupervised learning.
Machine learning--unsupervised Learning (unsupervised learning)
