DEPARTMENT OF TECHNOLOGY

SHIVAJI UNIVERSITY,
KOLHAPUR
Academic Year 2023-24
A
MINI PROJECT SYNOPSIS ON
“CRIME RATE PREDICTION”

Submitted To:
(Computer Science and Technology) Department
of Technology, Shivaji University, Kolhapur

Submitted By:
MS. AMRAPALI M. GAUTAM 22

MR. ADITYA B. GOSWAMI 27

MS. SRUSHTI M. CHAVAN 19

Under the Guidance of:

MR. H. A. TIRMARE
 CERTIFICATE

This is to certify that the project report entitled

" CRIME RATE PREDICTION "

Submitted To

Department Of Technology, Shivaji University, Kolhapur

has been completed under my guidance and supervision. To the best of my

knowledge and belief, the matter presented in this project report is original and
has not been submitted elsewhere for any other purpose.

Submitted by:

Sr. No. Name Roll. No.

1 AMRAPALI MILIND GAUTAM 22

2 ADITYA B. GOSWAMI 27

3 SRUSHTI MILIND CHAVAN 19

Project Guide Class Coordinator Program Coordinator

MR. H. A. TIRMARE MS.AMARJA BHOSALE DR. R.J. DESHMUKH

 ABSTRACT

Today, criminal activities are increasing worldwide due to modern

technologies, which have an adverse effect on society and common people.
The primary objective of this paper is to analyze the crime pattern and predict
the crime occurrence rate in the future to take appropriate measures against
it. One of the most suitable methods to apply to a huge amount of crime
datasets is the data mining technique to organize and analyze the data. The k-
means clustering algorithm is applied to form clusters based on crime in
various
regions to find generic patterns.

KEYWORDS: Crime , cyber security , K-Means algorithm

 INTRODUCTION

In the current scenario, criminals are becoming more technologically

advanced which serves as a major challenge for law agencies and government
officials to analyze huge amounts of criminal datasets. So, the basic
requirement is to
select appropriate crime analysis and data mining techniques. A data mining
algorithm can be used to get clusters of cities or areas having similar crime
occurrence rates and analyze the crime pattern. In this paper, the k-means
clustering technique of data mining is used to extract useful information from
the huge crime dataset. Clustering will be done based on the place where the
crime occurred. This will help to predict crimes that will occur in the future.
Which would help the police to identify the crime “hot spots” and take the
required actions
Classes of crimes are:
Robbery
Gambling
Accident
Violence
Murder
Kidnapping
This prediction, if put to good use, is of great help in investigating cases that
have happened. It can be used to suppress the crimes by installing some
measures if we know what type of crime is going to happen beforehand. This
will indirectly help reduce the rates of crimes and can help to improve security
in such required areas.
 LITERATURE REVIEW
[1] S.G. Krishnendu, P. P. Lakshmi, and L Nitha proposed a system to examine
the possibility of using clustering technology for crime analysis. They found
out which state has more or fewer criminals based on each cluster’s values.
The
result can be used to make various strategies for crime control in the future.
[2] Gouri Jha et al. proposed a system for finding a pattern that is necessary
for prediction and finding criminal patterns and behaviors. In the future, they
tend to focus on examining fuzzy clustering algorithms to apply to serial
killing-
related datasets and examine the analysis of the same.
[3] Khushabu A. Bokde et al. proposed a project that focuses on crime
analysis by considering crime homicide and plotting it concerning year. They
concluded that homicide is decreasing from 2009 to 2018.
 CHOICE OF TOPIC WITH REASONING

K-means excels at identifying groups (clusters) within data. Applied to crime

data, it can group areas with similar crime rates, demographics, or social
factors. This helps uncover patterns that might influence crime occurrence.
By clustering areas with high historical crime rates, K-means can flag zones
susceptible to future crimes. This allows law enforcement to allocate resources
more effectively for patrolling and preventive measures.

OBJECTIVE

• Optimize resource allocation for law enforcement: By pinpointing

potential crime hotspots, the application can guide law enforcement in
strategically deploying resources for better prevention and response.
• Enhance public safety: By informing users about areas with a
higher predicted crime rate, the application can empower them to
take
precautions and avoid risky situations.
PROPOSED SYSTEM
 METHODOLOGY

K-means Clustering:

 Define the desired number of clusters (k). This can be informed by prior
knowledge of crime patterns or through techniques like the elbow method.
 Initialize k centroids (random data points) that represent the center of each cluster.
 Assign each data point to the nearest centroid based on a distance metric (e.g.,
Euclidean distance).
 Re-calculate the centroid of each cluster based on the assigned data points.
 Repeat steps 4 and 5 until the centroids no longer change significantly
(convergence).

Analysis and Prediction:

 Analyze the characteristics of each cluster. Look for patterns in crime rates,
locations, and potentially socioeconomic factors associated with each cluster.
 Identify clusters with high historical crime rates. These areas are likely
candidates for future high crime rates.
 (Optional) Develop a scoring system based on cluster membership and other
relevant factors to estimate a potential crime rate for a specific area.

Visualization (Optional):

 Visualize the clusters on a map to identify spatial patterns of crime.

 CONCLUSION
In this system, the crime is analysed by implementing a k-means clustering
algorithm on the crime dataset. It will help the government officials easily
control the crime that occurs in a particular city. By grouping locations with
similar crime rates and potentially related social factors, K-means helps
uncover patterns that might influence crime occurrence. This allows Law
enforcement can prioritize patrolling and preventive measures in high-risk
areas identified by the clusters.

REFERENCES
1. Crime Analysis and Prediction using Optimized K-Means Algorithm

2. Criminal Behaviour Analysis and Segmentation using K-Means Clustering

3. Crime Detection Technique Using Data Mining and K-Means