Machine learning and azure ml studio

17
Machine Learning and Azure ML Studio Yogendra Tamang ASPNET Meetup 20 February 2016

Transcript of Machine learning and azure ml studio

Machine Learning and Azure ML Studio

Yogendra Tamang

ASPNET Meetup

20 February 2016

Outline

• Introduction

• Creating Models

• Regression

• Creating Models on Azure ML

• Demo

Machine Learning ?

• AI

• Learning Algorithm

• Lots of examples

• Testing and Evaluation

Machine Learning

• Machine Learning- Grew out of work in AI- New capability for computers

• Examples: - Database mining

• Large datasets from growth of automation/web. • E.g., Web click data, medical records, biology, engineering

- Applications can’t program by hand.• E.g., Autonomous helicopter, handwriting recognition, most of Natural Language

Processing (NLP), Computer Vision. - Self-customizing programs

• E.g., Amazon, Netflix product recommendations- Understanding human learning (brain, real AI).

Machine Learning

• Autonomous Helicopter

• Autonomous Driving

• Face Detection

Autonomous Cars, Facial Detection, NLP..

Azure ML

• Create Model• Get Data

• Pre-processing of data

• Define Features

• Train the Model• Choose and apply learning algorith

• Score and Test the model• Predict new automobile prices

Creating Models

1. Create new Experiment

2. Type in automobile to see Automobile Price Dataset1. Play Around with datasets

3. Pre-process Data

Getting data

• Automobile Price Data

• Each row for single automobile

Preprocessing Data

• Clean Missing Values• Normalized-Losses column Remove

• Remove any rows having missing data• Exclude normalized-loss[Use Project Columns]

• Clean rows having missing data [ Clean Missing Data Module]

Defining Features

• Requires experimentation and knowledge about context

• Some feature better at predicting target.

• Strong correlation with other features

Apply Learning Algorithm

• Classification or Regression ???

• Split Data to train and test

• Train [0.75] and test[0.25] … Use split data, Run Experiment

• Machine Learning -> Initialize Model -> Regression->Linear Regression

• Train Model Module

Training the model

• Train Model Module

• Left Port for Model, Right port for data

• Run experiment

Predict New Automobile Prices

• Score Model• Left Port from Train model

• Right Port from Test Data

• Run

• Evaluate Model

Thank you

References

• https://azure.microsoft.com/en-us/documentation/articles/machine-learning-create-experiment/