×
Community Blog Alibaba Cloud Machine Learning Platform for AI: Online Predictive Deployment for Health Monitoring

Alibaba Cloud Machine Learning Platform for AI: Online Predictive Deployment for Health Monitoring

This article describes how to use the Online Predictive Deployment feature of Alibaba Cloud's Machine Learning Platform for AI (PAI) to monitor user health in real time.

Join us at the Alibaba Cloud ACtivate Online Conference on March 5-6 to challenge assumptions, exchange ideas, and explore what is possible through digital transformation.

By Garvin Li

Models generated on the Alibaba Cloud Machine Learning Platform for AI (PAI) can be deployed online to generate APIs that can be invoked by other services. This document is based on the Heart Disease Prediction Case, and describes how to use the Online Predictive Deployment feature of the machine learning platform to monitor user health in real time.

1

Step 1: Model Deployment

Click Deploy in the lower section of the current experiment interface and select Online Predictive Deployment. Select the logistic regression model generated in the heart disease prediction case, as shown in the following screenshot.

2

3

Step 2: Model Deployment Information Configuration

Go to the model configuration page, as shown in the following screenshot.

4

Select a corresponding project. If you are using this for the first time, you need to enable the online prediction permission, which will be enabled in real time upon request. Set the number of instances occupied by the current model. Instances are described as follows.

  • By default, each project includes 30 instances. You can change the default value by opening a ticket. Deleting a deployed model will release instances of that model.
  • Instances determine a model's QPS. Each instance features 1 core and 2 GB memory.
  • The deployment limit of instances for a single model is [1,15].

Step 3: Model Management and Control

You can manage the deployed model in the interface shown below. Click View Service Details to check the newly deployed model.

5

You can manage the deployed model in Deployed Online Model, as shown in the following screenshot.

6

The following screenshot shows the model management interface. Different deployments of the same model can be differentiated by version. You can find the project where the model is located and the model name as shown in the following screenshot.

7

Step 4: Model Debugging

The model debugging page helps you understand the writing specifications of online prediction request parameters. Click Model Debugging shown in the preceding screenshot to access the model debugging page (Your browser may block this page. Please configure your browser to allow access to this link).

8

  1. The format of Request URL is "https://dtplus-cn-shanghai.data.aliyuncs.com/dataplus_261422/pai/prediction/projects/$projectname/onlinemodels/$modelname".
  2. Request body is a json format string. For example, for the logistic regression algorithm, information about each feature is required. Feature names need to match the corresponding feature names in the model list. The constant column doesn't require input.
    • dataValue: the value of the corresponding feature in the prediction set.
    • dataType: value type, as defined in the following screenshot.

      9

Step 5: Result Prediction

After the model is configured and debugged, edit the request body and send the request to get prediction results. Assume that parameters such as user gender, real-time blood pressure, and real-time heartbeat fluctuation are all 1. Send the following data to the server.

Here is the example body in this case.

1.    {
2.        "inputs": [
3.            {
4.                "sex": {
5.                    "dataType": 40,
6.                    "dataValue": 1
7.                },
8.                "cp": {
9.                    "dataType": 40,
10.                    "dataValue": 1
11.                },
12.                "fbs": {
13.                    "dataType": 40,
14.                    "dataValue": 1
15.                },
16.                "restecg": {
17.                    "dataType": 40,
18.                    "dataValue": 1
19.                },
20.                "exang": {
21.                    "dataType": 40,
22.                    "dataValue": 1
23.                },
24.                "slop": {
25.                    "dataType": 40,
26.                    "dataValue": 1
27.                },
28.                "thal": {
29.                    "dataType": 40,
30.                    "dataValue": 1
31.                },
32.                "age": {
33.                    "dataType": 40,
34.                    "dataValue": 1
35.                },
36.                "trestbps": {
37.                    "dataType": 40,
38.                    "dataValue": 1
39.                },
40.                "chol": {
41.                    "dataType": 40,
42.                    "dataValue": 1
43.                },
44.                "thalach": {
45.                    "dataType": 40,
46.                    "dataValue": 1
47.                }
48.            }
49.        ]
50.    }

Results will be returned after the request is sent. Returned results in this example show that the label is 1 (1 indicates illness and 0 indicates health) and that the illness probability is 0.98649974.

10

To learn how to invoke APIs, please visit https://help.aliyun.com/document_detail/30245.html.

0 0 0
Share on

GarvinLi

14 posts | 0 followers

You may also like

Comments