Rare Category Detection Using Hierarchical Mean Shift

1
(Rare) Category Detection Using Hierarchical Mean
Shift

• Pavan Vatturi (vatturi_at_eecs.oregonstate.edu)
• Weng-Keen Wong (wong_at_eecs.oregonstate.edu)

2
1. Introduction

• Applications for surveillance, scientific
  discovery and data cleaning require anomaly
  detection
• Anomalies often identified as statistically
  unusual data points
• Many detected anomalies are simply uninteresting
  or correspond to known sources of noise

3
1. Introduction
Known objects (99.9 of the data)
Anomalies (0.1 of the data)
Pictures from Sloan Digital Sky Survey
(http//www.sdss.org/iotw/archive.html) Pelleg,
D. (2004). Scalable and Practical Probability
Density Estimators for Scientific Anomaly
Detection. PhD Thesis, Carnegie Mellon
University.
Uninteresting (99 of anomalies)
Interesting (1 of anomalies)
4
1. Introduction

• Category Detection Pelleg and Moore 2004
  human-in-the-loop exploratory data analysis

Ask User to Label Categories of Interesting Data
Points
Data Set
Update Model with Labels
Build Model
Spot Interesting Data Points
5
1. Introduction

• User can
• Label a query data point under an existing
  category
• Or declare data point to belong to a previous
  undeclared category

Ask User to Label Categories of Interesting Data
Points
Data Set
Update Model with Labels
Build Model
Spot Interesting Data Points
6
1. Introduction

• Goal present to user a single instance from each
  category in as few queries as possible
• Difficult to detect rare categories if class
  imbalance is severe
• Interested in rare categories for anomaly
  detection

7
Outline

• Introduction
• Related Work
• Background
• Methodology
• Results
• Conclusion / Future Work

8
2. Related Work

• Interleave Pelleg and Moore 2004
• Nearest-Neighbor-based active learning for
  rare-category detection for multiple classes He
  and Carbonell 2008
• Multiple output identification Fine and Mansour
  2006

9
3. Background Mean Shift Fukunaga and Hostetler
1975
Reference data set
Mean shift vector (follows density gradient)
Query point
Center of Mass
Mean shift vector with kernel k
10
3. Background Mean Shift Fukunaga and Hostetler
1975
Reference data set
Convergence to cluster center
Query point
Center of Mass
11
3. Background Mean Shift Blurring
Reference data set
Query point
Center of Mass

• Blurring
• When query points are the same as the reference
  data set
• Progressively blurs the original data set

12
3. Background Mean Shift
End result of applying mean shift to a synthetic
data set
13
4. Methodology Overview

• Sphere the data
• Hierarchical Mean Shift
• Query user

14
4. Methodology Hierarchical Mean Shift
Repeatedly blur data using Mean Shift with
increasing bandwidth hnew k hold
15
4. Methodology Querying the User

• The data point closest to the cluster center is
  the representative data point. Rank
  representative data points for querying to user
  according to
• Outlierness Leung et al. 2000 for Cluster Ci

Lifetime of Ci Log (bandwidth when cluster Ci
is merged with other clusters bandwidth when
cluster Ci is formed)
16
4. Methodology Querying the User

• Rank representative data points for querying to
  user according to
• Compactness Isolation Leung et al. 2000 for
  Cluster Ci

17
4. Methodology Tiebreaker

• Ties may occur in Outlierness or
  Compactness/Isolation values.
• Highest Average Distance heuristic choose
  representative data point with highest average
  distance from user-labeled points.

18
5. Results
Data sets used in experiments
Shuttle, OptDigits, OptLetters, and Statlog were
subsampled to simulate class imbalance.
19
5. Results (Yeast)
Category detection metric queries before user
presented with at least one example from all
categories
20
5. Results
Number of hints to discover all classes
21
5. Results
Area under the category detection curve
22
6. Conclusion / Future Work

• Conclusions
• HMS-based methods consistently discover more
  categories in fewer queries than existing methods
• Do not need apriori knowledge of dataset
  properties

23
6. Conclusion / Future Work

• Future Work
• Better use of user feedback
• Presentation of an entire cluster to the user
  instead of a representative data point
• Improved computational efficiency
• Theoretical analysis