Search for answers or browse our knowledge base.
Analytics
0 out of 5 stars
5 Stars | 0% | |
4 Stars | 0% | |
3 Stars | 0% | |
2 Stars | 0% | |
1 Stars | 0% |
Sentiment Analysis Using Microsoft Azure Machine Learning Studio
Introduction
The Microsoft Azure Machine Learning Studio (MLStudio) is a workspace for crafting machine learning “experiments”. Experiments can be transformed in the MLStudio into Web Services, which allows them to be queried through a RESTful interface to get predications and classifications on demand.
To get started with MLStudio, visit https://studio.azureml.net/ and sign up for an account if you do not have one already. Microsoft currently allows users to deploy a limited number of experiments for free and provides 10gb of free hosting for your training data. If you aren’t familiar with MLStudio already, we recommend watching the tutorials and starting with some of the sample experiments or ones in the experiment gallery. Throughout this tutorial, we’ll use a Predictive Experiment for Twitter sentiment analysis found here: https://gallery.cortanaintelligence.com/Experiment/Predictive-Experiment-for-Twitter-sentiment-analysis-3. By the end of this tutorial, you will be able to query this experiment through a Vantiq Analytics Source to get the sentiment of any tweet.
Setting Up an MLStudio Experiment
After creating a MLStudio account and logging into the workspace, navigate to https://gallery.cortanaintelligence.com/Experiment/Predictive-Experiment-for-Twitter-sentiment-analysis-3 and click the “Open in Studio” link and click the checkmark on the “Copy experiment to gallery” pop-up. Open the experiment up from the list of experiments in your workspace, click “OK” on the bottom menu bar that indicates the experiment is being upgraded, then click run from the bottom menu. Once the run completes (can take around 10 minutes to train the model), select “Deploy Web Service” from the bottom menu and choose the recommended option.
Once you’ve deployed the web service, you should be taken to a page that specifies an API key and there should be a table below with rows for the Request/Response API and the Batch API. Select the blue “TEST” button from the Request/Response row and try it out, the response will show up at the bottom of the page. To see what API calls are actually being made here, click on Request/Response in the left-hand column. These docs, along with the API key from the previous page are all you need to request Sentiment Analysis from VAIL.
Configuring the Vantiq Analytics Source
In order to make requests to MLStudio for predictions using the predictive experiment, create a new Analytics Source in your Vantiq namespace with the name “Sentiment”. The Analytics Type is “Microsoft MLStudio” and the Request URI is the POST url specified in the Request/Response docs for your experiment. Specify the API key in the Access Token field then save the Analytics Source.
Using the Analytics Source in VAIL
Now that the Analytics Source has been configured, we can try it out in VAIL. Here’s an example procedure that takes a single parameter, a message to classify, and returns the response from running the predictive experiment:
PROCEDURE testSentimentOfStatement(msg)
// Construct the request body
// see the Request/Response docs for more details on how this is formulated
var payload = {"Inputs": { "input1": { "ColumnNames": ["tweet_text"], "Values": [[msg]]}}, "GlobalParameters": {}}
// Request the prediction from the source
var prediction = SELECT FROM SOURCE Sentiment WITH body = payload
log.info("Prediction result: {}", [prediction])
// Grab the result field
var result = prediction[0].Results
// Get the values from output1 (the name of the final activity in the experiment)
var output = result["output1"]
var values = output.value.Values
// values is actually an array, so get the first index
values = values[0]
// values is an array of length 2
// first value is the direction (positive, neutral, or negative)
// second value is the score
var polarity = values[0]
var score = values[1]
return {polarity: polarity, score: score}
0 out of 5 stars
5 Stars | 0% | |
4 Stars | 0% | |
3 Stars | 0% | |
2 Stars | 0% | |
1 Stars | 0% |
-
Getting Started
-
- Advanced Collaborations
- Analytics
- App Components
- Assemblies
- Catalogs Tutorial
- Client Builder
- Client Components
- Deployment Tutorial
- Floor Plan
- Introduction to Collaboration
- Natural Language Tutorial
- Sources
- Stateful Services
- System Modeler
- Testing the Debugging Tutorial
- Testing the Introductory Tutorial
- Testing the Source Tutorial
- User and Namespace Administration
- Show Remaining Articles ( 3 ) Collapse Articles
-
Product Documentation
-
-
-
- Accessing Namespaces in the Organization
- Active Resource Control Center
- Adding a New User to an Organization
- Adding a New User to the Application Namespace
- Administrators' Concepts and Terminology
- Authorizing Users to Access the Application
- Creating a Developer Namespace for the Organization Administrator
- Creating a New Application Namespace
- Creating Resources for New Namespaces
- Custom User Invites
- Deploying the GenAI Flow Service Connector
- Developer Tasks
- Handling Administrators Leaving
- Related Configuration
- Removing Namespace Administrators
- Self-Administration Tasks
- System Administration Tasks
- Viewing Lists of Users
- Show Remaining Articles ( 3 ) Collapse Articles
-
- Deploy Results Tab
- Deploying the same application to different environments
- Deployment
- Deployment Tool - Introduction
- Environment
- Environment Tab
- Node
- Project Partitions
- Redeploy On A Failed Node
- Reliable Deployment
- Settings Tab
- The Graph View
- The Tree View
- Undeploy
- Update Partitions
- Verify Application After Deployment
- Show Remaining Articles ( 1 ) Collapse Articles
-
- CheckedInsert/CheckedUpsert Command
- Command Line Options
- Delete Command
- Execute Command
- Export Command
- Find Command
- Help Command
- Import Command
- Insert Command
- Installation - Prerequisites
- Installation - Profile
- List Command
- Load Command
- Recommend Command
- Run Command
- Select Command
- Stop Command
- The Vantiq Command Line Interface (CLI) - Overview
- Upsert Command
- Show Remaining Articles ( 4 ) Collapse Articles
-
- App Execution Dashboard
- App With Split Dashboard
- Dashboard Navigation Bar
- Deprecated Dashboards
- Event Processing Dashboard
- General Dashboard Behavior
- Getting Started with Grafana
- Grafana Usage
- Monitoring Namespaces with Grafana
- Most Commonly Used Dashboards
- Namespace Monitoring Dashboards
- Organization Level Behavior
- Procedure and Rule Execution Dashboards
- Profiling Dashboards
- Reliable Event Dashboard
- Resource Usage Dashboard
- Service Execution Dashboard
- Service Handler Dashboard
- Source Activity Dashboard
- Storage Manager Dashboard
- Tensorflow Model Dashboard
- Type Storage Dashboard
- Show Remaining Articles ( 7 ) Collapse Articles
-
- Access to a Kubernetes Cluster
- Creating a K8s Cluster
- Delayed Processing
- Deploying K8s Installations to a Kubernetes Cluster
- Deploying the K8s Worker
- External Lifecycle Management Guide - Overview
- K8s Worker
- Kubernetes Components of a K8s Installation
- Kubernetes Namespaces
- Loading Images into a Kubernetes Cluster
- Managing K8s Installations
- Other Configuration Options
- System View
- Use of the self Cluster
- Using a Kubernetes Cluster
- Using Templates to Deploy the K8s Worker
- Vantiq Namespaces
- Verify Installation
- Show Remaining Articles ( 3 ) Collapse Articles
-
- Changing the System Password
- Creating a GenAIFlowService Service Connector
- Creating a New Organization and Namespace
- Deployment Methods
- Docker Deployment
- Edge Installation Management
- Edge Vision Server
- Executable JAR Deployment
- MongoDB
- Requirements
- Running the Vantiq Executable
- Setting the default LLMs API key
- Setting Up Vantiq Edge
- Vantiq Edge Reference Guide - Overview
- Vantiq Edge Self Node
- Windows bat file
- Show Remaining Articles ( 1 ) Collapse Articles
-
- Additional Buffer Semantics
- Applicability
- auditFrequency Quota
- Background
- Default Quotas
- Detailed Credit Quotas
- errorBreaker Quota
- errorReportingFrequency Quota
- Execution Credit Quota
- Execution Credit Quota - Diagnostics
- Execution Credit Quota - Mitigation
- Execution Rate Quota
- Execution Rate Quota - Diagnostics
- Execution Rate Quota - Mitigations
- executionTime Quota
- k8sResources Quota
- Quota Interactions
- receiveMessage Quota
- receiveMessage Quota - Diagnostics
- receiveMessage Quota - Mitigation
- reservedGroups Quota
- stackDepth Quota
- Stream Quota
- Terminology
- Workload Management
- Workload Management Conceptual Model
- Show Remaining Articles ( 11 ) Collapse Articles
-
-
-
-
- Advanced Use Cases
- Data Manipulation
- Defining Types
- Discovery from External Data Store
- Error Handling
- Installation and Use
- Native Language Implementation
- Restricting Capabilities
- Service Connectors
- Storage Manager Assembly Contents
- Storage Manager Service API
- Storage Manager Transactions
- Storage Managers - Introduction
- Transaction Support
-
-
-
- App Pane
- Autopsies
- Defining a Run Policy
- Defining a Test Suite - Properties
- Defining an Input
- Defining an Output
- Error Pane
- Integration Tests
- Populate Testing Namespace With Data
- Procedure Pane
- Rule Pane
- Running a Test in the IDE
- Running a Test through the REST Interface
- Source Mocking For Tests
- Unit Tests
- Vantiq Testing Reference Guide - Introduction
- Show Remaining Articles ( 1 ) Collapse Articles
-
-
-
- Assembly Configs
- Audits
- Catalog Members
- Catalogs
- Debug Configs
- Delegated Requests
- Documents
- Event Generators
- Groups
- Images
- K8s Clusters
- K8s Installations
- K8s Workitems
- LLMs
- Logs
- Namespaces
- Nodes
- Organizations
- Procedures
- Profiles
- Projects
- Resource Definition
- Resource Events
- Resource Relationship Model
- Resource Security Model
- Rules
- Scheduled Events
- Secrets
- Semantic Indexes
- Service Connectors
- Services
- Sources
- StorageManagers
- TensorFlowModels
- Test Reports
- Test Suites
- Tests
- Tokens
- Topics
- TrackingRegions
- Types
- Users
- Vantiq Resources
- Videos
- Show Remaining Articles ( 29 ) Collapse Articles
-
- Before Rules
- Built-In Services
- Data Manipulation
- Data Model Declarations
- Declaring Packages
- Defining Remote Connections
- Distributed Processing
- Error Handling
- Event Sending
- External State
- Flow Control
- General Use Procedures
- In-Memory State Management
- Iteration
- Logging
- Operators
- Package Scoping and Name Resolution
- Packages
- Packages
- Persistent State
- Procedure Execution
- Procedures
- PROCESSED BY Clause
- Resource Definition
- RETURN
- Rules
- Services
- Syntax
- Type Specific Procedures
- VAIL Declarations
- VAIL Types
- Variables
- Show Remaining Articles ( 17 ) Collapse Articles
-
-
-
Articles
-
- How To Video Shorts: Client Layouts
- How To Video Shorts: AI Functions
- How To Video Shorts: Analytics and ComputeStatistics
- How To Video Shorts: Calling Procedures by Properties
- How To Video Shorts: Client CSS
- How To Video Shorts: Invite Other Users to Your Namespace
- How To Video Shorts: SplitByGroup
- How To Video Shorts: The Vantiq API
- How To Video Shorts: The Vantiq IDE
- How To Video Shorts: The Vantiq Version Control System
- How To Video Shorts: Using Generative AI in Applications
- How-To Video Shorts: Managing AI Conversations
- How-To Videos: AI Design Model Assistant
- How-To Videos: AI Documentation Search
- Production Applications Best Practices