Identifying glass

1
Identifying glass

• Using the top-notch data-mining algorithms from
  the Leiden Institute of Advanced Computer Science
  (LIACS)
• Presented by Jan-Willem and Frans-Willem.

2
Why identifying glass?

• Crime scene glass found in leg of murdered
  person.
• Where did the glass come from?
• Might this have been the cause of death?

3
Research data Attributes

• Refractive Index (1.517865)
• Sodium (6.62)
• Magnesium (3.73)
• Aluminum (0.45)
• Silicon (76.38)
• Potassium (8.65)
• Calcium (2.76)
• Barium (0.40)
• Iron (1.01)

4
Decision Tree

• Barium lt 0.27
• (not relevant)
• Barium gt 0.27
• Silicon lt 70.16 build wind non-float
• Silicon gt 70.16 headlamps

5
Filling in our data

• Data Barium 0.40, Silicon 76.38
• Barium lt 0.27
• (not relevant)
• Barium gt 0.27
• Silicon lt 70.16 build wind non-float
• Silicon gt 70.16 headlamps

6
Headlamps

• Run down?!
• Hit n Run?

7
Possible suspects

• Jan-Peter- Has no drivers license- No previous
  criminal records
• Bea- No longer valid to drive because of age-
  Is always escorted by motors, can not run down
  persons.
• Peter- Has just got his drivers license-
  Left-headlight duct taped!

8
Using our data

• Jan-Peter- Has no drivers license- No previous
  criminal records
• Bea- No longer valid to drive because of age-
  Is always escorted by motors, can not run down
  persons.
• Peter- Has just got his drivers license-
  Left-headlight duct taped!

9
Conclusion

• After investigating the car of Peter, and taking
  samples of his headlights, we found out that it
  matched the exact percentages as the glass
  particle found in the victims leg.
• This was enough to convince the judge, and get
  him behind bars for his lifetime!