5 Machine Learning Job Guide

Workflow for creating a classification algorithm job

---

1. Go to Phoenix Layer2 Computation

2. Enter the job list page

3. Click on Create a new job

4. Input job name

5. Choose a computation type

Choose decision tree, logistic regression or svm as computation type

6. Click on ‘Upload a Document’

Upload feature data, target data and testing data(optional) separately. The file must be in .csv format and can only contain regular characters such as letters, numbers, and underlines. The feature data and testing data should have the same number of columns.

The sample data is as follows:
feature data, target data, testing data.

7. Click on Submit

8. Click on Check

Enter the job list page and click the check button.

9. View the result

Wait for a while, The training model and testing results will be uploaded to ipfs, you can access them by clicking the links.

Workflow for creating a clustering algorithm job

---

1. Go to Phoenix Layer2 Computation

The first four steps are the same as creating a classification algorithm

2. Choose a computation type

Choose a clustering algorithm from birch clustering, k-means clustering, spectral clustering

3. Upload the training feature dataset

The clustering algorithm only needs to upload the feature dataset

4. Click on Submit

5. Click on Check

Go to the job list page to find the job you just created, and click the check button

6. View the result

Wait for a while, the trained model and clustering results will be uploaded to the blockchain. You can access them by clicking the links