Incapacitation, Recidivism and Predicting Behavior

1
Incapacitation, Recidivism and Predicting Behavior

• Easha Anand
• Intro. To Data Mining
• April 24, 2007

2
Background

• Crime Control Act of 1984 and USSC
• Idea in U.S. is deterrence, rather than
  punishment
• Tending toward formulaeUSPC in D.C. uses 14
  variables
• U.S. prison pop. topped 2 million,
  parole/probation topped 7 million

3
Strategies for Incapacitation

• Charge-based
• Historically the case most USSC guidelines
• Selective
• USPC and D.C. Code offendersbased on
  individuals characteristics
• New research focuses on criminal career and
  predicting patterns therein (participation,
  frequency, seriousness, length, patterning)

4
Rationale

• The tendency is toward objective decision-making
  processes to improve accuracy.
• More and more variables codified as we can track
  offenders.
• Sophistication of statistical methods used to
  combine predictors seems to be relevant to
  outcomes.

5
The Dataset

• 6,000 men incarcerated in the 1960s, chosen at
  random
• Collected life history info, official
  institutional record, inmate questionnaire,
  psychological tests
• 26 years later, followed up with Bureau of
  Criminal Statistics
• Offenses characterized along six dimensions
  Nuisance, physical harm, property damage, drugs,
  fraud, crimes against social order
• Used 4,897 records

6
Dataset (contd)
7
Problems With Data

• Dichotomous dependent variable for behavior?
• Purging potential bias
• Done after age 70 OR
• When 10 years arrest-free
• No record of out-of-state crimes

8
Philosophical Problems

• Metric for success
• False positives 30,000 arrests could have been
  prevented!
• False negatives 1,413 people jailed
  unnecessarily
• Reduced crime could have to do with repentance,
  increased policing, age, etc. and not with
  incapacitation at all

9
Data Pre-processing

• Only used records where had both 1962 and 1988
  data
• Priors of previous convictions weighted by
  severity of crime
• PriorsP of previous periods of incarceration
  weighted by length
• Inst_(M,P,V,F,etc.) of arrests weighted by
  severity of crime in each of six categories

10
of Arrests to Desistance (R2 .159)
11
of Arrests to Desistance (Violent Crimes
Onlyn1,998)
R2 .061 plt.05
12
What Next?

• Multiple Linear Regression
• Try using different things as classnuisance
  only, arrest rate, crime-free time
• Try different predictorshave 119 variables
• BUT
• No reason to believe predictors are linearly
  independent
• No reason to believe non-linear correlation

13
What Next?

• Better technique Decision trees
• White Box model mimics human decisionmaking
• Use some kind of feature-selection algorithm?
• Maybe ensemble learning, once feature-selection
  is in place?

14
Acknowledgements

• Trevor Gardner, UC Berkeley
• Don Gottfredson, Rutgers University
• Bureau of Criminal Statistics