Download Presentation slide - Big Data Analytics Nigeria

DATA MINING APPLICATION IN CRIME
ANALYSIS AND CLASSIFICATION
Obuandike Georgina N. 1. John Alhasan 2, M. B. Abdullahi 3
1. Department of Mathematical Sciences and IT, Federal University, Dutsinma,
Katsina state, Nigeria.
2,3. Department of Computer Science, Federal University of Technology, Minna,
Niger State, Nigeria
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
Introduction
• What is data mining
• Data Mining and Data Warehouse: a close ally, takes
80% of time.
• Big Data: it come more quickly and in different
format, it is characterized by 3Vs, volume, variety and
velocity
• Big Data Analytic: This is the application of advanced
analytic techniques to very big data sets.
• Place of data mining in crime analysis
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
1
Research Questions
• Is data mining able to unravel hidden patterns from
the crime data
• Is the technique selected appropriate for the analysis
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
2
Methodology
Figure 1: Work Methodology Flowchart
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
3
Evaluation Metrics
• Sensitivity: It is a statistics that shows the records that are
correctly labelled by the classifier.
• Specificity: It is simply a report of instances incorrectly
labelled as correct instances
• Precision: Simply measures exact relevant data retrieved.
High precision means the model returns more relevant data
than irrelevant data.
• Kappa: measures the relationship between classified
instances and true classes. It usually lies between [0, 1], the
value of 1 means perfect relationship while 0 means random
guessing.
• Accuracy: this shows the percentage of correctly classified
instances in each classification model
• Time: Implies time taken to perform the
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
4
Exploratory Analysis Result
Figure 2: Association of age and educational qualification
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
5
Exploratory Analysis Result cont.
Figure 3: Association of offence and educational qualification
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
6
Classification Result
Evaluation Metrics
Time
Accuracy
TP Rate
FP Rate
Kappa
Precision
Recall
ROC curve
NB
0.05 secs
93
0.935
0.067
0.8696
0.935
0.935
0.989
C4.5
1.06 secs
97
0.971
0.027
0.9409
0.971
0.971
0.986
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
7
Conclusion
Data mining has the capability that makes it simple
convenient and suitable for data extraction from large
databases.
It employs different mining algorithms for it’s work. Many
agencies gather data for its operational purposes, such
data can be mined to discover some relevant patterns
that can aid in decision making.
The analysis of crime data will help to unravel crime
pattern and nature of those who commits such crime so
that appropriate strategies and rules will be put in place
to control such crimes.
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
8
Conclusion cont
The work reveals that the majority of the inmates that
commit crime are between the ages of 18 to 34 and have
low or no educational qualification and are either self
employed or not doing anything at all.
The classification result reveals that 98 percent of these
groups of people are threat to the society.
Thus, the researchers are of the opinion that government
should encourage education and our youths should be
gainfully employed.
3rd Big Data Analytics and Innovation Conference, 22-25 November, NDC, Abuja, Nigeria
9