1© 2017 The MathWorks, Inc.

Machine Learning for Financial

Applications

Dr. Alexander Diethert, Application Engineer

2

Challenges for Analysts

How to work with your data to .

… analyze for insights?

… identify patterns?

… develop predictive models?

…deploy to collaborators?

… innovate?

Opportunities

3

Customer Examples

4

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

5

Computational Finance Workflow

Research and Quantify

Data Analysis

& Visualization

Financial

Modeling

Application

Development

Reporting

Applications

Production

Share

Automate

Files

Databases

Datafeeds

Access

6

Machine LearningCharacteristics and Examples

Characteristics

– Lots of data (many variables)

– System too complex to know

the governing equation(e.g., black-box modeling)

Examples

– Pattern recognition (speech, images)

– Financial algorithms (credit scoring, algo trading)

– Energy forecasting (load, price)

– Biology (tumor detection, drug discovery)

93.68%

2.44%

0.14%

0.03%

0.03%

0.00%

0.00%

0.00%

5.55%

92.60%

4.18%

0.23%

0.12%

0.00%

0.00%

0.00%

0.59%

4.03%

91.02%

7.49%

0.73%

0.11%

0.00%

0.00%

0.18%

0.73%

3.90%

87.86%

8.27%

0.82%

0.37%

0.00%

0.00%

0.15%

0.60%

3.78%

86.74%

9.64%

1.84%

0.00%

0.00%

0.00%

0.08%

0.39%

3.28%

85.37%

6.24%

0.00%

0.00%

0.00%

0.00%

0.06%

0.18%

2.41%

81.88%

0.00%

0.00%

0.06%

0.08%

0.16%

0.64%

1.64%

9.67%

100.00%

AAA AA A BBB BB B CCC D

AAA

AA

A

BBB

BB

B

CCC

D

7

Challenges – Machine Learning

Significant technical expertise required

No “one size fits all” solution

Locked into Black Box solutions

Time required to conduct the analysis

8

Overview – Machine Learning

Machine

Learning

Supervised

Learning

Classification

Regression

Unsupervised

LearningClustering

Group and interpretdata based only

on input data

Develop predictivemodel based on bothinput and output data

Type of Learning Categories of Algorithms

9

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

10

Unsupervised Learning

Clustering

k-Means,

Fuzzy C-Means

Hierarchical

Neural

Networks

Gaussian

Mixture

Hidden Markov

Model

11

Unsupervised Learning

Clustering

k-Means,

Fuzzy C-Means

Hierarchical

Neural

Networks

Gaussian

Mixture

Hidden Markov

Model

12

ClusteringOverview

What is clustering?

– Segment data into groups,

based on data similarity

Why use clustering?

– Identify outliers

– Resulting groups may be

the matter of interest

How is clustering done?

– Can be achieved by various algorithms

– It is an iterative process (involving trial and error)

-0.1 0 0.1 0.2 0.3 0.4 0.5 0.60

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

13

Example – Clustering Corporate Bonds

Goal:

– Cluster similar corporate bonds

together

Approach:

– Cluster the bonds data using distance-

based and probability-based

techniques

14

Example – Clustering Corporate Bonds

Numerous clustering functions with

rich documentation

Interactive visualizations to aid

discovery

Viewable source; not a black box

Rapid exploration & development

15

Classification Learner App

16

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

17

Supervised Learning

Regression

Non-linear Reg.

(GLM, Logistic)

Linear

RegressionDecision Trees

Ensemble

Methods

Neural

Networks

Classification

Nearest

Neighbor

Discriminant

AnalysisNaive Bayes

Support Vector

Machines

18

Supervised Learning - Workflow

Known data

Known responses

Model

Train the Model

Model

New Data

Predicted

Responses

Use for Prediction

Measure Accuracy

Select Model

Import Data

Explore Data

Data

Prepare Data

Speed up Computations

19

ClassificationOverview

What is classification?

– Predicting the best group for each point

– “Learns” from labeled observations

– Uses input features

Why use classification?

– Accurately group data never seen before

How is classification done?

– Can use several algorithms to build a predictive model

– Good training data is critical

-0.1 0 0.1 0.2 0.3 0.4 0.5 0.60

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Group1

Group2

Group3

Group4

Group5

Group6

Group7

Group8

20

Example: Calibrating a rating system

Key tasks1. Perform ad hoc data analysis (linear regression)

2. Improve existing analysis (logistic regression)

3. Create a robust classifier: TreeBagger

ID WC_TA RE_TA EBIT_TA MVE_BVTD S_TA Industry

60644 0.049 0.220 0.041 2.400 0.489 6

Ratings

AA

Rating System

21

Regression Learner App

22

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

23

Neural Network Use Cases

Function Approximation and Nonlinear Regression

Pattern Recognition and Classification

Clustering– Self-Organizing Maps

– Competitive Layers

Time Series and Dynamic Systems– Modeling and Prediction with NARX and Time-Delay Networks

– Creating Simulink Models

24

Neural Network Apps

25

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

26

Share Programs Outside of MATLAB

Deploy your MATLAB code to people who don’t need MATLAB

27

Application Author

End User

1

2

Sharing Standalone Applications

MATLAB

ExcelAdd-in Hadoop

StandaloneApplication

Toolboxes

MATLAB Compiler

MATLAB

Runtime3

28

MATLAB

MATLAB

Compiler SDK

C/C++ExcelAdd-in JavaHadoop .NET

MATLAB

Compiler

MATLABProduction

Server

StandaloneApplication

Which Product will Fit Your Needs?

MATLAB Compiler for sharing MATLAB programs without integration

programming

MATLAB Compiler SDK provides implementation and platform flexibility for

software developers

MATLAB Production Server provides the most efficient development path

for secure and scalable web and enterprise applications

Python

29

The Range of Application Platforms

30

MATLAB Production Server

Web

Server

Application

Server

Database Server

Warehouse

Application 2

Application 3

Application 1

MATLAB Production Server

MATLAB

Component

Web Applications

Desktop Applications

Trade Execution;

Batch Applications

31

Agenda

Introduction

Unsupervised Learning: Clustering Bond Data

Supervised Learning: Create Rating System

Neural Network Based Time Series Forecasting

Outlook: Distributing MATLAB Applications

Summary

32

Challenges MATLAB Solution

Time (loss of productivity) Rapid analysis and application developmentHigh productivity from data preparation, interactive

exploration, visualizations.

Extract value from data Machine learning, Video, Image, and FinancialDepth and breadth of algorithms in classification, clustering,

and regression

Computation speed Fast training and computationParallel computation, Optimized libraries

Time to deploy & integrate Ease of deployment and leveraging enterprisePush-button deployment into production

Technology risk High-quality libraries and supportIndustry-standard algorithms in use in production

Access to support, training and advisory services when

needed

MATLAB for Machine Learning

33

Learn More: Machine Learning with

MATLABmathworks.com/machine-learning

34

Thank you!