32
Christian Schaiter Categorization of Snow Profile Data into Predefined Class Templates Academic advisors: Univ.-Prof. Dr. Günther Specht DI Robert Binna MASTER THESIS Leopold-Franzens-University Innsbruck Institute of Computer Science

Categorization of Snow Profile Data into Predefined Class Templates

Embed Size (px)

DESCRIPTION

Today avalanche research primarily focuses on the reliable prediction of avalanche descents by using meteorological data and snowpack properties. This thesis complements these approaches by providing means to automatically categorize snow proles into ten predened templates, which may be used as an indicator whether potential avalanches might carry o deeper layers of snow to form avalanches of higher magnitude. For this purpose, the snow hardness prole is taken and transformed into a symbolic internal data format based on weighted strings. This representation allows for the application of well-known methods such as string alignments, which provide the foundation for the classication system. Furthermore, as a secondary strategy, the class-templates themselves are modied in a neat way that does not distort their overall shape, thereby providing matching candidates for a larger portion of input proles. Altogether, the proposed system converts the input prole into the internal representation, takes each of the modied template versions and determines a proper alignment of hardness layers as the basis for a nal error score computation that enables an ordering among the contemplable template types. The work in this thesis may be generalized to approach a wider range of problems and is closely related to the eld of time series data mining.

Citation preview

Page 1: Categorization of Snow Profile Data into Predefined Class Templates

Christian Schaiter

Categorization of Snow

Profile Data into

Predefined Class

Templates

Academic advisors: Univ.-Prof. Dr. Günther Specht

DI Robert Binna

MASTER THESIS

Leopold-Franzens-University Innsbruck

Institute of Computer Science

Page 2: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 2

Content

Introduction and motivation 1

What you may expect in the next 15 minutes...

Overview of the developed system 2

Alignment of hardness profiles 3

Conclusion 4

Page 3: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 2

Overview

Introduction and motivation 1

Current chapter: Introduction

Alignment of hardness profiles 3

Conclusion

Overview of the developed system 2

4

Page 4: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 3

Introduction and motivation

Master thesis was done in cooperation with the Tyrolean

avalanche warning service (TAWS)

Goal: Find a way to

determine the potential

magnitude of an

avalanche

Aim of the thesis

Page 5: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 4

Introduction and motivation

TAWS frequently takes snow profiles from different areas

Most relevant property of a snow profile for this thesis:

Hardness profile

Snow profiles, hardness profiles, template types

Sn

ow

heig

ht

[cm

]

Snow hardness

Hardness

profile

Page 6: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 5

Introduction and motivation

Some example hardness profiles (collected from the TAWS):

Snow profiles, hardness profiles, template types

Page 7: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 6

Introduction and motivation

Idea: Introduce a group of predefined hardness profile types

(template types, class templates)

These template types describe the overall composition of the

snowpack

They may be used to

estimate the magnitude

of avalanches

10 types defined so far

Snow profiles, hardness profiles, template types

Profile types (class templates)

Page 8: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 7

Introduction and motivation

Snow profiles, hardness profiles, template types

Hardness profile

Determine type

Goal: Automatically find the associated profile type for

each collected hardness profile

Reject hardness profile if no reasonable profile type is found

Profile types (class templates)

Snow hardness

Sn

ow

heig

ht

[cm

]

Page 9: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 8

Overview

Overview of the developed system

Introduction and motivation 1

Current chapter: Overview of the developed system

2

Alignment of hardness profiles 3

Conclusion 4

Page 10: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 9

Overview of the developed system

Schematic view of the proposed classification system

Page 11: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 10

Overview

Introduction and motivation 1

Current chapter: Alignment of hardness profiles

Conclusion 4

Overview of the developed system 2

Alignment of hardness profiles 3

Page 12: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 11

Alignment of hardness profiles

Hardness profiles may be viewed as

time series (clockwise rotated by 90°)

Height corresponds to time axis

Hardness values map to the amplitude

dimension

Desired properties of an internal data

format:

Height independence

Shape preservation

Local height warping

Representation format of hardness profiles

90°

Page 13: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 11

Alignment of hardness profiles

Hardness profiles may be viewed as

time series (clockwise rotated by 90°)

Height corresponds to time axis

Hardness values map to the amplitude

dimension

Desired properties of an internal data

format:

Height independence

Shape preservation

Local height warping

Representation format of hardness profiles

matching

Page 14: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 12

Alignment of hardness profiles

Idea: Use symbolic representation (weighted strings)

A weighted string is composed of weighted characters

Direction character:

D (down), U (up)

Height (as a percentage)

Hardness difference

(among adjacent layers)

Compact form uses only

direction characters

Hardness profiles as weighted strings

6 5 4 3 2 1

3050 ,.,D

110 ,.,D

50350 .,., D

610 ,.,U U

D

D

D

U

D

D

D

53050 .,.,U

51150 .,., D

1150 ,.,D

51050 .,., D

Page 15: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 13

Alignment of hardness profiles

Hardness profiles are aligned based on their compact strings

Weights are required to calculate a penalty score

Goal: Find a global alignment with a maximum number of

matching characters (optimal alignment)

Edit transcript describes operations (match M, deletion D,

insertion I) required to transform the first string into the second

Principle of hardness profile alignment

The compact string UUDUD and DUDUU may be aligned as

– U U D – U D U U D U D – –

D – U D U U – or – – D U D U U , etc I D M M I M D D D M M M I I

String alignment examples

Page 16: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 14

Alignment of hardness profiles

Principle of hardness profile alignment

Well-matching profile Badly-matching profile

U D D – U U D D D U D D U U D D D

U D D D – U D D D U – D – – – – –

M M M I D M M M M M D M D D D D D

Page 17: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 15

Alignment of hardness profiles

Calculation of optimal alignments

Algorithms for calculating optimal string alignments:

Page 18: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 16

Alignment of hardness profiles

Calculation of optimal alignments

Page 19: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 17

Alignment of hardness profiles

After all optimal alignments have been found: Compute penalty

scores

2 types of penalties:

Match-penalty

for differences in:

- Height

- Hardness

Mismatch-penalty

for mismatching blocks

- Much more severe

Computation of penalty scores

match-penalty

Page 20: Categorization of Snow Profile Data into Predefined Class Templates

Intrusion Detection Systems for SOA 18

Overview

Introduction and motivation 1

Current chapter: Conclusion

Overview of the developed system 2

Conclusion 4

Alignment of hardness profiles 3

Page 21: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 19

Conclusion

Template types are used to

estimate magnitude of avalanches

Classification is based on string

alignment techniques

Well matching strings similar

hardness profiles

Achieved success rate of > 90% for

example set (if not rejected)

Approach may be used for any time

series data

What you have heard in the last 15 minutes

Page 22: Categorization of Snow Profile Data into Predefined Class Templates

Any Question?

Thank you for your attention!

Page 23: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 21

Backup slides

Distance measures: Cosine of angle, Euclidean Distance

Similarity measure: Cosine of angle

Dissimilarity measure: Euclidean Distance

Euclidean Distance:

Take the smallest distance

between points

in the vector space

Cosine of angle:

Take the smallest angle

between vectors

Page 24: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 22

Backup slides

Problems with the Euclidean Distance

Consider the hardness profile

(in blue) and its dedicated

template (in green)

In principle they are

well-matching

Nevertheless, a large error

occurs (in red) when applying

the Euclidean Distance

measure

ED does not handle “height

warps”

Page 25: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 23

Backup slides

Calculation of optimal alignments

Compute a distance table (with dynamic programming)

Example: Compact strings UUDUD and DUDUU

Page 26: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 24

Backup slides

Calculation of optimal alignments

Based on the distance table, compute the edit graph and

perform a traceback

U U D U D – –

– – D U D U U

D D M M M I I

– U U D – U D

D – U D U U –

I D M M I M D

Page 27: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 25

Backup slides

Template type duplication

Template type versioning:

Create different versions with multiple gradation steps

Page 28: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 26

Backup slides

Template type duplication

Variations of template type versions:

Scale up one layer at the cost of the other layers

Page 29: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 27

Backup slides

Profile matching

Alignment of hardness layers: Limitations

Example should be of type 4, but perfectly matches type 3

Page 30: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 28

Backup slides

Profile matching

Alignment of hardness layers: Limitations

Problem of deceptive local extremes

Causing false negatives

badly matching

perfect match

Page 31: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 29

Backup slides

Profile matching

Alignment of hardness layers: Limitations

Problem of abusing misalignments

Causing false positives

Apparently wrong profiles may achieve too good penalty scores

(see examples)

Page 32: Categorization of Snow Profile Data into Predefined Class Templates

Categorization of Snow Profile Data into Predefined Class Templates 30

Backup slides

Genetic Algorithm principles

Mechanics of a simple Genetic Algorithm: 5 Steps

Step I: Encoding of the search domain with a small alphabet

Step II: Creation of an initial string population

Step III: Reproduction of strings

Step IV: Crossover of strings

Step V: Mutation of strings