Alibaba Cloud PAI-Studio
The course is part of this learning path
This course introduces visual development through PAI. We will give a detailed introduction to the basic concepts and features of PAI-Studio. And conduct a practical exercise of a simple machine learning experiment based on PAI-Studio for you to understand it better.
This is a practical demonstration for the binary classification algorithm of the Support Vector Machine by using PAI console. You need to open the website of PAI console firstly. And then, click Model Training and select Studio Modeling Visualization. The right side is a project list. At the first time, you should create a project if you don't have any project.
When you already created a project, you could click Machine Learning, which your project matches with. And now, you come in a new website, which name is Machine Learning Platform for AI. When coming in platform, the first thing you need to do is creating a new experiment. You should click on the icon of Home to the left most column, then click on this button and select New Experiment. Type your experiment's name and description, and choose where your project saved to in your own folder. When everything is okay, click the OK button.
The next step, what you need to do is constructing your own dataset. You should go to the graphical interface of data by clicking on the Data Source icon. Then, click the Create Table. In this interface, you can fill in your data set name in the column of Table Name. And the lifecycle means how long the data set you upload can stay in the PAI platform. If you modify 180 to one, it means the data set which you upload will be destroyed after 24 hours. The following parts will show how to fill in schema. These two files are the datasets which you want to upload.
The data onto these data sets are formatted. The first column is the ID column in every dataset. And obviously, you have 10 items in every data set. The second column is the Labels column of the dataset. In general, the label of the dataset is one or minus one, which means positive or negative for the binary classification. The rest of the columns are features columns. The other dataset is same to this dataset. You could also upload the dataset similarly. Back to platform, you need to fill this table correctly.
Every column in that dataset puts into one-to-one relationship with one line in this table, sequentially. And the type of every line must be right and do not select the option of partition column. When finished all, click the Next button. We should modify the row delimiter and column delimiter. For these two datasets, the row delimiter is /n and the column delimiter is /t. After modification, you should upload the datasets by the button of Select File. In the end, click OK button. And if there's not any error message in the top right hand corner, one data set is uploaded successfully.
The other data set needs to be done almost by rote. After complete the operation of uploading datasets, you should drag the dataset to canvas. Next, you need to build your operation logic. You click the Components icon and drag some components to canvas. The module of the Logistic Regression for binary classification is in the Binary Classification folder, which belongs to the machine learning. And the module of Prediction is in the machine learning as well.
The module of Binary Classification Evaluation and Confusion Matrix belong to Evaluation folder, which is in Machine Learning. Other modules can also be found easily in the left list. If you cannot found a module, you could use the searching function. According to the logic in modules, you need to build such a logic diagraph. You should point the one module of dataset to SVMs model and point these two to Prediction. And in accordance with the logic of evaluation, point the Prediction to two modules of evaluation. There are some attributes you need to modify.
Click the module of the Logic Regression for binary classification, and feel in all these contents correctly. You need to select the right columns for training feature and target. In addition, you also need to form the settings of binary classification evaluation by selecting label column. To the current location, we finish all we need to do to build a binary classification algorithm of the Support Vector Machine. And in the end, we should click Run button and the program will start at once.
When the project is finished, you can check the answers to the tests set by clicking the note of prediction with the right and selecting View Data. In this table, you can see the true results and prediction results. And the prediction score shows what the trend of the sample being predicted is. The larger the absolute value is, the more credible it is. Meanwhile, you can check the data's in others nodes by using View Data. The practical demonstration for the binary classification algorithm of the Support Vector Machine is finished.
Alibaba Cloud, founded in 2009, is a global leader in cloud computing and artificial intelligence, providing services to thousands of enterprises, developers, and governments organizations in more than 200 countries and regions. Committed to the success of its customers, Alibaba Cloud provides reliable and secure cloud computing and data processing capabilities as a part of its online solutions.