Functional Data Analysis

Jiguo Cao, PhD

Canada Research Chair in Data Science
Director of the Pacific Blue Cross Health Informatics Laboratory
Associate Professor, Department of Statistics and Actuarial Science
Associate Faculty Member, School of Computing Science
Simon Fraser University, Vancouver, Canada

Functional Data Analysis : 1

Classical Functional Data
Measures of position of nib of a pen writing ”fda”. 20 replications, measurements taken at
200 hertz.

Functional Data Analysis : 2

What is Functional Data?

20 replications, 1401 observations within replications, 2 dimensions

Functional Data Analysis : 3

What is Functional Data?

20 replications, 1401 observations within replications, 2 dimensions

Functional data is often complex:

Functional Data Analysis : 3

What is Functional Data?

20 replications, 1401 observations within replications, 2 dimensions

Functional data is often complex:

• often a large number of related

quantities

Functional Data Analysis : 3

What is Functional Data?

20 replications, 1401 observations within replications, 2 dimensions

Functional data is often complex:

• often a large number of related

quantities

• viewing each replication as a single

observation can make the data easier to
think about (once we have the right
machinery)

Functional Data Analysis : 3

What is Functional Data?

What are the most obvious features of these data?

Functional Data Analysis : 4

What is Functional Data?

What are the most obvious features of these data?

• quantity

Functional Data Analysis : 4

What is Functional Data?

What are the most obvious features of these data?

• quantity

• frequency (resolution)

Functional Data Analysis : 4

What is Functional Data?

What are the most obvious features of these data?

• quantity

• frequency (resolution)

• similar trends

Functional Data Analysis : 4

What is Functional Data?

Most important: smoothness

Functional Data Analysis : 5

What is Functional Data?

Most important: smoothness

These data describe (nearly) a process that

changes smoothing, and continuously over
time.

Functional Data Analysis : 5

What is Functional Data?

Most important: smoothness

These data describe (nearly) a process that

changes smoothing, and continuously over
time.
Functional Data Analysis = Analysis of data
that are functions.

Functional Data Analysis : 5

What is Functional Data?

Most important: smoothness

These data describe (nearly) a process that

changes smoothing, and continuously over
time.
Functional Data Analysis = Analysis of data
that are functions.
Domain is usually time, but can be anything:
space, energy ...

Functional Data Analysis : 5

What is Functional Data?
20 replications

Functional data analysis involves repeated measures of the same process.

Functional Data Analysis : 6
What is Functional Data?

20 replications, 1401 observations within replications

Functional data is often complicated:

• not easily described by mathematical

formulae

• variation between replications even

harder to describe

Functional Data Analysis : 7

Characteristics

• Data are measurements of smooth processes over time

• We usually do not want to make parametric assumptions about those processes.

• Often have multiple measurements of the same process

• We are interested in describing the variation of processes.

• Frequently, collected data have high resolution and low noise.

• Can be applied to any estimate of a smooth process.

Functional Data Analysis : 8

About Functional Data Analysis

1. FDA is New
• First named in Dalzell & Ramsay, 1991
• Relatively little penetration into applied fields (= easy publication)
• Several competing methodologies (we focus on one)
• Limited public software/resources
• data analysis rather than inference

2. Functional Data is Complex

• Requires more thought/judgement than a t-test
• data needs pre-processing
• parametric inference is rarely available/appropriate

Functional Data Analysis : 9

What will this course offer?
Audience: application areas with functional data

Focus: • What can Functional Data Analysis do?

• How do I make it happen?

Software: packages in R, Matlab

Goals: Enabling you to

• Understand and interpret the result of FDA applied to real data
• Use existing FDA libraries to analyze functional data
• Evaluate its usefulness/correctness
• Extend the methods in existing software if you need to

Not Covered: reproducing-kernel Hilbert spaces, asymptotics, theorems...

Functional Data Analysis : 10
Pre-requisites and Recommendations

Pre-requisites: multiple linear regression

Useful: Life will be easier if you do not need to learn some of the following:

• R/Matlab or other programming experience

• Calculus
• Matrix algebra
• Multivariate statistics
• Computational statistics

Any necessary material will be covered in class, but will be out of context.

Functional Data Analysis : 11

Resources
Textbook: Ramsay and Silverman, 2005, Functional Data Analysis, Springer.

Books: • Ramsay and Silverman, 2002, Applied Functional Data Analysis, Springer.
• Ramsay, Hooker and Graves, 2009, Functional Data Analysis with R and
MATLAB, Springer

Online: • https://github.com/caojiguo/FDAcourse2019/
All lecture slides, exercises, computer codes, etc will be posted here.
• https://www.youtube.com/channel/UC1Wh20PhCEOnrEkk58WweFA
Recorded lectures will be posted here.
• http://www.functionaldata.org for FDA
• http://www.r-project.org a general site for R
Functional Data Analysis : 12
Functional Data Analysis : 13
Back to "What is Functional Data"

Or What isn’t Functional Data?

Do my data need to look this good?

Functional Data Analysis : 14

Data may be measured more noisily

We need to find the smooth process under the data.

Functional Data Analysis : 15
Data may be measured more sparsely

• Data are low noise but low-resolution

• Measured at unequal intervals

• We know that the curves must always

increase

Functional Data Analysis : 16

MIT Growth and Development Study

40
● • 162 girls
●
● ●
●
● ●

• 6.4 measurements on average (sparse)

● ●
●
●
●
●

● ●
● ●
30 ● ● ●
●
Body Fat Percentage

● ●
●
● ●
● ●
●
●

●
● • irregular
● ● ● ●
●
● ●
● ●
● ●
● ● ●
● ●
●
● ●
20 ●
●
●
●
● ●
●
●

●
●

●
●
●
10

−2.5 0.0 2.5

Time Relateve to Menarche

Functional Data Analysis : 17

Our goal

• Estimating the underlying functions

Functional Data Analysis : 18

Our goal

• Estimating the underlying functions

• Estimating the derivative functions

Functional Data Analysis : 18

For instance...
id=150 id=160 id=27 id=55
10
Estimated Derivatives

−5

−2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0
Time Relative to Menarche
id=150 id=160 id=27 id=55
Body Fat Percentage

40
● ●
●
●
●
30 ●
● ●
●
● ● ●
● ●
●
20 ●
●
● ●
●
●
10 ●

−2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0 −2.5 0.0 2.5 5.0
Time Relative to Menarche

Functional Data Analysis : 19

Necessities for Functional Data

• must believably derive from a smooth process

• process should not be easily parameterizable (should not be able to write down a formula)

• enough data to resolve the essential features of the process (peaks, zero-crossings,
speed... will depend on application)

• some repetition in the process

• do not need equally-spaced or perfect measurements

Functional Data Analysis : 20

Common Sources

medical monitoring: EEG, ECG, fMRI, blood pressure ...

medical tests: HIV antibodies, flu screens...

biology: animal behavior (whale songs, fly egg-laying...)

environmental monitoring: weather, pollution, solar radiation, traffic ...

optotrack experiments: psychology/physiology

economics/marketing: macro-trends, futures markets

web data: e-bay auction prices, google trends

Functional Data Analysis : 21

Essential Questions

Or what can FDA do for me?

• How do we go from discrete to functional data?

• How do we describe random variation in functional data?

• How do we decide if groups of functional data are different?

• How do we relate functional data to other data? To other functional data?

• What is special about functional data?

• Aligning functions (registration)

• Use of rates of change (dynamics)

Functional Data Analysis : 22

Approximate Class Agenda

1. From data to functional data

• Basis expansions and smoothing
• The fda library
• Positive and monotone smoothing

2. Exploring Functional Data

• Means, variances, covariances
• Functional PCA

3. Functional Linear Models

Functional Data Analysis : 23