TITLE: An Improved Crime Analytics and Prediction System for Expert

Recommendation using Data Mining techniques

INTRODUCTION:
With the boom in population, the number of crime occurrences has increased
exponentially. The proposed system aims to overcome the problems faced by the police
department mainly on area wise crime analysis, prediction of criminal accused, and
recommend suitable experts(police officer) based on offense occurred. The Analytical
system works based on the data provided by the police department. The data set contains
accused details, victim details, location, type of crime, and handled officer details. An
Expert is recommended based on the location and their area of expertise. The proposed
analytical system also provides us an insight into the crimes occurring in different
locations. The system analyzes criminal records, the system can suggest a list of possible
criminals.

OBJECTIVE:
 To improve the accuracy of crime detection and prediction by expert recommendation
using advanced data analytical methodologies.

METHODOLOGY:
In this section, we are going to present a conceptual model of our system and explain the
various aspects of it in a systematic manner.

A. Criminal Dataset Acquisition and Creation

The crime is huge in number and hence they must be stored into various groups by
considering various parameters. The government of India has categorized crimes into many
groups.
They are classified as:
 Violent Crime
A violent crime, or crime of violence, is a crime in which the offender uses or threatens to use
violent force upon the victim. This entails both crimes in which the violent act is the objective,
such as murder, as well as crimes in which violence is the means to an end, (including
criminal ends) such as robbery. Violent crimes include crimes committed with and without
weapons.
 Organized Crime
A set of large criminal organizations (often competing for markets and territories) that deals in
illegal goods and services.

 Property Crime
Property crime is a category of crime that includes, among other crimes, burglary, theft, motor
vehicle theft and shoplifting. Property crime only involves the taking of money or property,
and does not involve force or threat of force against a victim.
All the criminals fall into one or more of these categories and therefore they can be
predicted. The details about the criminals and the location are gathered from the officer through a
form which contains related questions. The questions are framed in such a way that it covers all
details about the criminals and the handled officers. This form will be filled by the officer by
using a system or smart phone through websites and application. In case of non-availability of
computer or smart phone, the officer can submit the information through the government centers.
All the information’s are collected and stored in a centralized database, and this data is used to
for further analysis. Before proceeding to the algorithm, the data will be generally sorted by
breaking them into various groups based on parameters like criminal records, locality, etc. The
sorted data will be further sorted into sub groups by enforcing more specific parameters during
the time of requirement.
The system being proposed will hold an enormous amount of data and the uniqueness is
maintained throughout the process. In India, the criminal reports are maintained by the police
department and hence the data about the criminals, officer's, the crime occurred in that area is
stored. This data will also help us to identify individual’s location.

FIGURE 1.FLOW DIAGRAM OF CRIME ANALYTICS AND PREDICTION

SYSTEM
 B. Problem Solving using Decision Tree for Crime Analytics
We have the data about various criminals and their records and now we need an algorithm to
predict the criminals so the best-suited algorithm is Decision tree algorithm. Decision tree
learning uses a decision tree as a predictive model, to go from observations about an item
(represented in the branches) to conclusions about the item's target value (represented in the
leaves). It is one of the predictive modeling approaches used in statistics, data
mining and machine learning. Tree models where the target variable can take a discrete set of
values are called classification trees; in these tree structures, leaves represent class labels and
branches represent conjunctions of features that lead to those class labels. Decision trees where
the target variable can take continuous values (typically real numbers) are called regression trees.
Basic Algorithm:
 Tree is constructed in a top down recursive divide and conquer manner
 At start, all the training examples are at the root
 Attributes are categorical(if continuous they are discretized in advance)
 Input data is partitioned recursively based on selected attributes
 Test attributes at each node are selected on the basis of a heuristic or statistical measure

Conditions for stopping partitioning:

 All samples for a given node belong to the same class
 There are no remaining attributes for further partitioning – majority voting is employed
for classifying the leaf
 There are no samples left

The Decision tree representation of our proposed crime analytics system is shown in Figure
2.The root node has 3 branches, such as violent crime, organized crime, and property crime and
the leaf node represents the class labels of crime analytical system .

FIGURE 2.DECISION TREE REPRESENTATION FOR CRIME ANALYTICS

 C. Expert Recommdetation and Prediction using Case Based Reasoning
 A case-based reasoner solves new problems by using or adapting solutions that were
used to solve old problems. It offers a reasoning paradigm that is similar to the way
many people routinely solve problems. The case based recommendation of experts
representation is shown in Figure 3.

FIGURE 3: CASE BASED REASONING FOR RECOMMENDATION OF EXPERTS

 The expert details along with the crimes handled will be stored in the case base. The type
of crime will be given as input query, the pattern matching will be done with the case
base and the matching cases with expert details will be retrieved for recommendation. if
no cases matches the current case, new case will be evaluated and hence the expert
officer will be given as result and it will be retained in the case base for further use

WORKFLOW
Work Plan DURATION
START (days)
END DATE Task Name
DATE

10/1/17 11/1/17 Data Collection 30

Database Design and

11/2/17 12/2/17 30
Entry

12/3/17 12/14/17 Data Verification 11

12/15/17 1/20/18 Server Design 35

Implementation
1/21/18 2/17/18 26
Architecture
2/18/18 3/8/18 Interface Architecture 20

3/9/18 3/23/18 Testing 14

3/24/18 3/25/18 Deployment 1

6/14 8/3 9/22 11/11 12/31 2/19 4/10 5/30

Data Collection

Database Design and Entry

Data Verification

Server Design

Implementation Architecture

Interface Architecture

Testing

Deployment

BUDGET:

S.No Requirements Cost

1 Server 3000
2 Web Hosting 5000
3 SSL Encryption 2000