www.ijcrt.

org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

CRIME ANALYSIS AND PREDICTION USING

DATAMINING: A REVIEW
1Shanjana A.S, 2Dr.R.Porkodi
1PGStudent(Department of computer Science), 2Associate Professer(Department of Computer Science)
1Bharathiar University,
2Bharathiar University

ABSTRACT

Crime analysis and prediction is a systematic approach for identifying the crime. This system can predict
region which have high probability for crime occurrences and visualize crime prone area. Using the concept
of data mining we can extract previously unknown, useful information from an unstructured data. The
extraction of new information is predicted using the existing datasets. Crimes are treacherous and common
social problem faced worldwide. Crimes affect the quality of life ,economic growth and reputation of nation.

With the aim of securing the society from crimes, there is a need for advanced systems and new approaches
for improving the crime analytics for protecting their communities. We propose a system which can analysis,
detect, and predict various crime probability in given region. This paper explains various types of criminal
analysis and crime prediction using several data mining techniques.

KEYWORDS

Crime prediction, Decision trees, Linear Regression, k-means.

1. INTRODUCTION

Day by day crime data rate is increasing because the modern technologies and hi-tech methods are helps the
criminals to achieving the illegal activities .according to Crime Record Bureau crimes like burglary, arson etc
have been increased while crimes like murder, sex, abuse, gang rap etc have been increased [2].crime data
will be collected from various blogs, news and websites. The huge data is used as a record for creating a
crime report database. The knowledge which is acquired from the data mining techniques will help in
reducing crimes as it helps in finding the culprits faster and also the areas that are most affected by crime
IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 465
 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
[11].Data mining helps in solving the crimes faster and this technique gives good results when applied on
crime dataset, the information obtained from the data mining techniques can help the police department.

A particular approach has been found to be useful by the police, which is the identification of crime ‘hot
spots ‘which indicates areas with a high concentration of crime [1].Use of data mining techniques can
produce important results from crime report datasets. The very step in study of crime is crime analysis. Crime
analysis is exploring, inter relating and detecting relationship between the various crimes and characteristics f
the crime. This analysis helps in preparing statistics, queries and maps on demand. It also helps to see if a
crime in a certain known pattern or a new pattern necessary.

Crimes can be predicted as the criminal are active and operate in their comfort zones. Once successful they
try to replicate the crime under similar circumstances. The occurrences of crime depended on several factors
such as intelligence of criminals, security of a location,etc The work has followed the steps that used in data
analysis, in which the important phases are Data collection ,data classification, pattern identification,
prediction and visualization. The proposed framework uses different visualization techniques to show the
trends of crimes and various ways that can predicts the crime using machine learning algorithm.

2. CRIME DATA ANALYSIS

Collection and analysis of crime related data are imperative to secure agencies. the use of a coherent methods
to classify these data based on the rate and location of occurrences, detection of the hidden pattern among the
committed crimes at different times, and prediction of their future relationship are the most important aspects
that have to be addressed. One of the most popular approaches is hot spot analysis. Some of the most popular
approaches used for this purpose of point pattern analysis and clustering/distances statistics. Another popular
approach is the discovery of pattern or trends through various techniques from data mining, text mining and
spatial analysis, and self-organizing maps.[1]An crime analysis tool should be able to identify crime patterns
quickly and in an efficient manner for future crime pattern detection and action.

The main purpose of crime analysis is:

 Extraction of crime pattern by crime analysis and based on available criminal information.
 Crime recognition [3].
 Problem of identifying techniques that can efficient and accurate.

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 466

 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
s2.1 CRIME ANALYSIS METHODOLOGY

The crime analysis methodologies are:-

 Data Collection
 Classification
 Pattern Identification
 Prediction
 Visualization

Data Collection

Classification

Pattern Identification

Prediction

Visualization

Crime analysis steps

Data collection

The data collection is first methodology in crime analysis. Data’s are collected from various different
websites, news sites and blogs. The collected data is stored into database for further process. This is
unstructured data and it is object oriented programming which is easy to use and flexible.

Crime data is an unstructured data since no of field, content, and size of the document can differ from one
document to another the better option is to have a schema less database. Also the absence of joins reduces the
complexity. Other benefits of using an unstructured database are that:

 Large volume of structured, semi-structured, and unstructured data.

 Object-Oriented programming that is easy to use and flexible.

Classification

In this step use Naive Bayes Algorithm which is supervised learning method. Naive Bayes classifier is a
probabilistic classifier which when given an input gives a probability distribution of set of all classes rather
than providing a single output. One of the main advantages of the Naïve bayes Classifier is simple, and
coverage quicker than logistic regression [2].Compare to other algorithm like SVM (Support Vector
IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 467
 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
machine) which takes lots of memory.

Using naïve Bays algorithm is create a model by training crime data related to vandalism, murder, robbery,
burglary, sex abuse, gang rape, etc. Naive Bayes is that works well for small amount of training to calculate
the classification parameter. Estimating probability sometimes while checking a probability P(A) * P(B/D)
*P(C/D) * P(E/D) where P(C/D)=0[2].

Pattern Identification

A third step is the pattern identification where we have identify trends and patterns in crime. For finding
crime pattern that occurs frequently we are using apriori algorithm. Apriori can be used to determine
association rule which highlight general trends in the database. By using pattern identification it will helps to
the police officials in an effective manner and avoid the crime occurrences in particular place by providing
security, CCTV, fixing alarms etc.
Crime Prediction

The second Approach is predicting the crime type that might occur in a specific location within particular
time. To predict an expected crime type is provide four related features of the crime. The features are:
occurrence month, the occurrence day of the week, the occurrences time and the crime location. Prediction is
stating probability of an event in future period time. A Classification approach is used crime prediction in
data miningn[1]classify areas into hotspots and cold spots and to predictive an area will be a hotspot for
residential burglary. Variety of classification techniques are used for predicting the crime:-[1]

 K-Nearest Neighbor (k-NN)

 Decision trees (J48)
 Support Vector Machine (SVM)
 Neural Networks
 Naïve Bayes and ensemble learning

Linear Regression methods are also used for predicting the crime prediction. Based on the crime probability.
The formula for a regression line is

Y=aX + b where, Y is the predicted score, b is the slope of the line, and A is the Y intercept. b = r sx\sy

And the intercept (A) can be calculated as A=MY –bMX.

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 468

 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
Some Theories are used to predicting the crimes are:

 Integrated theory
 Biological theory
 Psychological theory
 Sociological theory
 Conflict theory
 Victimization theory
 Choice theory

Visualization

The crime prone area can be graphically reoresented using a heat amp which indicates level of activity,dark
colour indicates low activity and brighter colour indicates the high activity.

Advantages of using heat map are [2].

 Numerical and category based color images

 Gradient color range
 Analyze only the data we want
 Out of range data is automatically discarded.

Map showing crime prone area

FLOW CHART OF CRIME ANALYSIS AND PREDICTION

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 469

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

Take crime dataset

Filter dataset according to

requirements

Open Rapid miner tool and read

excel file of crime dataset

Apply replace missing value operator

and excute

Perform normalization operator on

resultant dataset and execute

Perform k means clustering on

resultant dataset and excute

Perform plot view and get cluster

Perform crime analysis on cluster

formed

Flow chart of crime analysi

Crime database

Prediction Estimate
Algorithm d Result

Output result

Test data

Predicting future crime trends [3]

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 470
 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

3. DATA

This dataset contains a record of incidents that the Austin Police Department responded to and wrote a report.
Data is from 2003 to present. This dataset is updated weekly. Understanding the following conditions will
allow you to get the most out of the data provided. Due to the methodological differences in data collection,
different data sources may produce different results. This database is updated weekly, and a similar or same
search done on different dates can produce different results.

Comparisons should not be made between numbers generated with this database to any other official police
reports. Data provided represents only calls for police service where a report was written. Totals in the
database may vary considerably from official totals following investigation and final categorization.
Therefore, the data should not be used for comparisons with Uniform Crime Report statistics. The Austin
Police Department does not assume any liability for any decision made or action taken or not taken by the
recipient in reliance upon any information or data provided. Pursuant to section 552.301 (c) of the
Government Code, the City of Austin has designated certain addresses to receive requests for public
information sent by electronic mail.

Summary of the crime Report dataset Rows: 2.21M

Columns: 27

In the dataset contains different types of crimes (attributes) are considered like murder, rape, kidnapping,
dacoit, robbery, burglary, cheating, dowry deaths, arson etc.

No Crime type No of records

1 Anti social behavior 44,070

2 Burglary 22,081

3 Violent crime 21,333

4 Other theft 19,538

5 Vehicle crime 18,260

6 Drugs 8,336

7 Roberry 2,166

8 Theft from person 799

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 471

 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

Types of crimes [1]

4. ALGORITHMS

Our experiment choose the algorithm are

 Instance based algorithm

 Decision tree
 Linear regression
 K-means algorithm

1. Instance Based Algorithm

-The instance based algorithm is also called as tge machine based learning is a family of learning algorithm
that, instead of performing explicit generalization, compares new problems instances with instance seen in
training, which have been stored in memory. These stored their training set when predicting a value or class
for a new instances, they compute distance training instances to make a decision.

The algorithm in this category for numerical prediction can divided into two types: similarity- based, e.g.,
Euclidean or entropy based and regression-based e.g., LWL Since regression is one of the most popular
methods for numerical prediction[1].

The advantages of the Instances based Algorithm is it over other methods of machine learning is its ability to
adapt its model of machine learning is its ability to adapt its model to previously unseen data. Instance based
learners may simply store a new instance or throw an old instance away. The Disadvantages of the instances
based Algorithm are its need more storage and computational complexity.

2. Linear Regression

-It is simple form of regression. Linear regression attempts to model the relationship between the two
variables by fitting a linear equation to observe the data. this is widely used in statistics. For this purpose
,linear functions are used for which the unknown parameter i.e., weight of the independent variables, are
estimated from the training data[1].this can be used to predict the values One of the most common estimating
method is least mean square.

Linear regression algorithms for predicting include simple regression multiple regression and pace
regression, which is suitable for data of high dimensionality and only accepts binary nominal attributes.[1].

The main advantages of the linear regressions is gain a far greater understanding of the variables that can
impact its success in the coming weeks,months and years into the future.The disadvantages of the regression

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 472

 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
is its linearity.If the data has non linear dependencies, a linear regression model will ouput the best fitting
line which may not fit very well.

3. Decision Tree

Decision tree is used for both the prediction and classification. for the classification purpose a function can
be learned this is intervals defined by splits on the individuals attributes value

Area Nota VIP Crim Crime

Sensiti ble Pres Inal
vity event ence group
Yes Yes Yes No Yes

Yes Yes No Yes No

No No No Yes No

Yes No No No No

Yes Yes Yes Yes Yes

No Yes No No No

TABLE: Decision Tree for Delhi [2]

Attribute 1
Yes No

Attribute 2 Attribute 4

No Yes Yes No
Attribute 3 No Yes No

Yes
Yes No
Decision tree for the above table
A Root node, that has incoming edges and zero or more outgoing edges.

 Internal nodes, each of which has one incoming edges and two or more outgoing edges.
 Leaf node or end node, each of which has exactly one incoming edge and no outgoing edges.
For prediction purpose, the decision trees algorithm for classification have been adapted to output a
numerical value the main difference

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 473

 www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
Is that the leaves of the tree have numerical values, unlike classification trees have class labels.

Advantages of the decision trees are It is very simple to understand and help determine worst,best and
expected values for different scenarios.it can be combined with other decision techniques.Some of the
Disadvantages of the Decision tree are They are unstable, They are often relatively inaccurate, Calculation
can get vey complex.

4. K-Means Algorithm

:K –means is the simplest and most commonly used portioning algorithm among the clustering algorithm in
scientific and industrial software[3].Acceptance of k means is mainly due to its being simple .This algorithm
is also suitable for clustering of a large datasets since it has much less computational complexity grows
linearly by increasing of the data points.

Advantages of the k-means algorithm are relatively simple to implement, Scales to large dataset, Guarantees
convergence, easily adapts to new examples. Disadvantages of the k-means algorithm are Choosing
manually, Being dependent on initial values, clustering data of varying sizes and density.

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 474

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
LITERATURE SURVEY
S NTITL E AUTHO R JOURNA L OVERVIEW OF METHO ABSTRA ACCURAC FUTURE
O PAPER DS USED CT THEAM Y RESULTWORK
OBTAINE
D
1 An Ginger Internatio Crime data has Crime This Paper The Further
Explor Saltos and nal Journal been systematically prediction, explore experiments experiment
ation of Minhaela of recorded by the Data mining, models were can be
Crime Cocea Informatio n police for many Open data, predicting the conducted conducted to
Predict and years and in the last regression frequency of using the investigate
ion Technolog decades, there has , decision several SCIAMA other aspects
Using y& Decision been a surge of tress, types of High such as
Data Making Open Crime instance- crimes. in Performanc e time frame
Minin g (World Data and of apps or based this paper Computer for
on Scientific web based learning. using three Cluster at prediction,
Open ) application types of the University the amount of
Data displaying crime algorithms of Portsmouth data
statistics on maps, they are and the necessary for
both by official instance- Weka reliable
sources, such as based software prediction
from police UK, learning, models, and
and other sources Decision predictive
usingthe same trees, models for
official data. In regression. particular
this paper types of
investigate crime.
predict many types
of crime and discuss
their applicability.

2 Crime Shiju First According to theNaïve bayes, In this This paper Our system
Analys isSathyade Internatio Crime record Apriori paper predict has tested predicts
and van,Deva n nal Bureau crimes likealgorithm, the crime the accuracy crime prone
Predict M.S, Conferenc e burglary, arson, etcDecision based on of classificatio regions in
ion Surya on have been decreasedtree, NER, the n and India on a
Using Gangadh networks & while crimes likeMongo Db, occurrence s prediction particular
Data aran soft murder, sex, GraphDBs of the based on day. It will
Minin computing abuse, gang rap crimes and different test be more
g

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 475

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

etc have been can visualize sets. accurate if we

increased. In this crime prone Classificatio n consider a
paper data’s are areas. is done particular
collected from Using the based on the stage/regio
various sources like concept of Bayes theorem n. Also
websites, blogs, data mining which showed another
news sites, social can extract more than problem is
media, RSS feeds previously 90% that we are
etc. This huge data unknown, accuracy. not predicting
is used as a record useful the time in
for creating a crime information which the
record database. from an crime is
The crime unstructure d happening .
analysis using five data.
steps they are Data
Collection,
Classification,
Pattern
identification,
prediction, and
Visualization.

3 Crime Khushab u Internatio This paper is mainlyCrime, The main From the From the
Detect A.Bokde, nal Journal focusing on crime Clustering objective of accuracy clustered
ion Tisksha P. of Engineeri Analysis, Clustering , K-Means this result, crime result it is
Techni Kakade, ng Research and Algorithm paper is to data mining easy to
ques Dnyanes & Clustering by K- classify has a identify
Using hwari S. Technolog means algorithms clustered promising crime trends
data Tumsare, ieEngineer methods. some of crimes based future for over year and
Minin g Chetan G. ing the purpose of on increasing the can be used to
and Wadhai (IJETER) crime analysis are occurrence effectivenes s design
K- B.E Extraction of frequently and precaution
Means Student crime patterns by during efficiency of methods for
crime analysis different criminal and future.
and based on years. Data intelligence
available criminal mining is analysis .
information, crime used to
recognition. extensively in.
Clustering means terms of
analysis,

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 476

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

Division of a set of investigaga

data or objects to a tion and
number of clusters. discovery of
There by cluster is patterns for
composed of a set occurrence s
of similar data of crime.
which have same as
a group. K- means
is the
simplest and most
commonly sued
portioning
algorithm among
the clustering
algorithms in
scientific and
industrial software.

4 Survey Benjamin ICTACT Criminology isCriminolo The The In future

on crime Fredrick Journal on process that is usedgy, Crime extraction of quantitative ,there is a
analysi s David. H Soft for identify the crime Analysis, the new analysis plan for
and and computing characteristics and Crime information is produced applying
predict Suruliand i crime Prediction predicted results which other
ion using identify. By using , Data using the shows the classificati on
data this data mining Mining existing increase in algorithm on
mining algorithms will be datasets. the Accuracy the
techni able to produce The criminals level of crime data
ques. crime reports and can also be classificatio n and
help in the predicted because of improving
identification of based on using the the accuracy
criminal much the crime GA to in prediction.
faster than any data. The optimize the
human could. The main aim parameters.
criminals when of this
leaving the crime work is to
scene does leave perform a
some trace which survey on
can be used as the

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 477

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

clue to identify the supervised

criminals learning and
.This process is used unsupervis ed
to learning
identifying the techniques
criminals based on that has
the clues or been applied
information given towards
by the local peoples. criminal
in this paper using identificati
some criminal on.
analysis methods
they are Text,
content and NPL-
based
methods, Crime
patterns .
5 Crime Tushar Indian Crime analysis is K-Means, Preventing Result will The future
Pattern Sonawan Journal of exploring, inter Cluster, are taken to be in the scope is
Analys ev, Shirin Computer relating and Correlatio n reduce the form of prediction of
is, Shaikh,S Science and detecting increasing correlation the
Visual haista Engineeri ng relationship number of between crime will be
ization s Shaikh, (IJCSE) between various cases of various crime displayed
And Rahul crimes and crime against and using various
Predict Shinde, characteristics of women. A location of diagram pie
ion Asif crimes. The main huge amount crime. crime charts,
Using Sayyad. aim of this of can also be heat maps,
Data research paper data set is correlated on spikes and
Minin g consists of generated the basis of agegraphs.
developing every year on ,group,
analytical data the basis of location of
mining methods reporting the crime &
that can crime. type of
systematically crime.
address the
complex problems
related
to various form of
crime. From this
literacy study, it

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 478

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

could be
concluded that
crime details
increasingly to very
large
quantities running
into zota bytes.
6 Crime Rajkumar Internatio Main contribution Data mining, Crime Analyzed and As a future
Analys is.S, nal Journal of this paper is to machine hotspot compared extension of
And Sakkarai of recent propose a new learning, prediction different our
predict Pandi.M, trends in approach based crime has algorithm on work, we
ion using Soundary a engineerin g on deep learning analysis, previously crime plan to
Data Jagan.J,V and success in crime been data determine apply more
Minin g arnikasre reaserch different prediction. suggested. which classificati on
Techni e.P classification task Crime algorithm models
ques such object hotspot performs to increase
detection image prediction better of crime
recognition natural leverages crime prediction
image past data in prediction. accuracy and
processing and order to to
dimensionality identify enhance the
reduction deep crime overall
learning algorithms hotspots, or performanc e.
use social media
deep architectures of data.
multiple layers to
extract features fr
om raw data.

7 Syste Sapreet Internatio This paper is Crime data Crime data The results In future to
matic kaur,Dr. nal Journals explains techniques mining, mining has of this result improve the
Revie w Williamj of Advanced used, crime data the ability may help performanc e
of eet Singh Research in challenges analysis, of extracting new potential of these
Crime Computer addressed, systematic useful users in classificati
Data Science methodologies used, review, information understandi ng on. Hence,
Minin g (IJARCS) and crime data systematic and hidden the range of the usage
mining and analysis study patterns from available of other
paper. The the crime data classificati
methodologies is large data mining on

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 479

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

composed of three sets. The technologies algorithm

stages the first stage crime data and like machine
involves the mining technologies learning will
research work challenges . be
related to crime data are becoming explored in
mining, second stimulating future.
stage is concerned opportuniti es
with for the
establishing a coming years.
classification and
the third stage is
involves the
presentation of
summary of
research in crime
data mining and
analysis and
report of this
survey

8 Survey Ayishesh Internatio Different Crime This paperA Survey is In future

paper on im nal Journal classification prediction, investigates aconducted so predicts what
Crime Almaw,K of Pure and algorithm in naïve Bayes, number that kind
problem resolving
Predict alyani Applied J48,Artific of data crime of crime
are nominated
ion using kadam Mathemati c based on the ial network mining forecasting might occur
Ensem suitable algorithms can be next in a
ble requirements in and ensemble improved by particular
Appro crime data learning the use of district within
ach prediction.One which are efficient data a
technique may applied on collection and certain
provide better
crime data data periods of
accuracy value
than different mining. mining time and
techniques which This paper strategies. identifier the
are nominated for delivers season
solving the reasonable and time
particular problems investigatio n factor at
.Some of data which crimes
papers introduced
mining are
combined
different models techniques occur more
to achieve better

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 480

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

performance which and ensemble frequently

overcome classificati on happening
the individual techniques crime.
models to achieve
for discovery
better performance
which overcomes and
the individual prediction of
models called upcoming
ensemble learning. crime.

9 Revie w Dr .M Internatio Clustering Data mining, Extract By observing In future

on Sreedevi, nal Journal techniques is used machine previously this survey more data
Crime A.harsha of to existing and learning, unknown this is a mining
Analys is vardhan Innovation known crimes. It is crime useful theoretical techniques
and reddy, Researchi ng able to detect newer analysis, information study for are used for
Predict Ch.venka ta Science, and crime from an several predicts the
ion Sai Engineeri ng unknown patterns in prediction unstructure d methods and crime pattern.
Using Krishna and future .this data. This methodologi
Data Reddy Technolog y paper collect the paper es in
Minin g dataset from explains identificatio n
Techni crime data various types of crime and
ques available in of criminals
national crime criminal which
bureau of records. analysis and includes
by using the crime text/NLP,
correlated data set prediction crime patterns,
we can identify the using several geo location,
correlated crimes. data mining prisoner
And the k means techniques methods,
algorithm is used communicat
for group the data ion based
into same methods data
characteristics. collection,

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 481

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882

classificatio n,
pattern
identificatio n,
prediction.

10 Crime Deepika Internatio According to this Clustering The approach A goodThe future
analysi s k,k, Smitha nal journal paper burglary , consists of accuracy ofenhanceme nt
Journa l Vinod of Engineeri and robbery has classificati the following 99.93% is of this
of ng & reduced over a on, steps – data obtained and research
engine Technolog y period of 53 tears Visualizati preprocessi this work focuses
ering & by79.84%and on, k- ng, verifies the on training
techno 28.85% means, clustering, correctness of bots to
logy respectively. crime random classificati the predict the
like murder and forest, on, and instances. crime prone
kidnapping Neural visualizatio area by
has hiked by networks n. Data using
7.39% and mining machine
47.80% techniques learning
respectively. Crime are often techniques.
analysis is a applied to
part of criminolog y
criminology plays as it
an important role in provides
crime detection good results.
.data mining helps
in solving the
crimes faster and
this
techniques give
good results when
applied on crime
dataset, the
information
obtained from the
data mining
techniques can
help the police
department.

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 482

www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
CONCLUSION

In this paper focused on building predictive models for crime frequencies per crime type per month. The
crime rates in India are increasing day by day due to many factors such as increase in poverty,
implementation, corruption, etc. The proposed model is very useful for both the investigating agencies and the
police official in taking necessary steps to reduce crime. The project helps the crime analysis to analysis these
crime networks by means of various interactive visualization.

Future enhancement of this research work on training bots to predict the crime prone areas by using machine
learning techniques. Since, machine learning is similar to data mining advanced concept of machine learning
can be used for better prediction. The data privacy, reliability, accuracy can be improved for enhanced
prediction.

REFERENCE

[1] Ginger Saltos and Mihaela Coacea, An Exploration of Crime prediction Using Data Mining on Open
Data, International journal of Information technology & Decision Making ,2017.

[2] Shiju Sathyadevan, Devan M.S, Surya Gangadharan.S, Crime Analysis and Prediction Using Data
Mining, First International Conference on networks & soft computing (IEEE) 2014.

[3] Khushabu A.Bokde, Tisksha P.Kakade, Dnyaneshwari S. Tumasare, Chetan G.Wadhai B.E Student,
Crime Detection Techniques Using Data Mining and K-Means, International Journal of Engineering Research
& technology (IJERT) ,2018

[4] H.Benjamin Fredrick David and A.Suruliandi,Survey on crime analysis and prediction using data mining
techniques, ICTACT Journal on Soft computing, 2017.

[5] Tushar Sonawanev, Shirin Shaikh, rahul Shinde, Asif Sayyad, Crime Pattern Analysis, Visualization And
prediction Using Data Mining, Indian Journal of Computer Science and Engineering (IJCSE), 2015.

[6] RajKumar.S, Sakkarai Pandi.M, Crime Analysis and prediction using data mining techniques,
International Journal of recent trends in engineering & research,2019.

[7] Sarpreet kaur, Dr. Williamjeet Singh, Systematic review of crime data mining, International Journal of
Advanced Research in computer science , 2015.

[8] Ayisheshim Almaw, Kalyani Kadam, Survey Paper on Crime Prediction using Ensemble Approach,
International journal of Pure and Applied Mathematics,2018.

[9] Dr .M.Sreedevi, A.Harha Vardhan Reddy, ch.Venkata Sai Krishna Reddy, Review on crime Analysis and
prediction Using Data Mining Techniques, International Journal of Innovative Research in Science
Engineering and technology ,2018.
IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 483
www.ijcrt.org © 2021 IJCRT | Volume 9, Issue 2 February 2021 | ISSN: 2320-2882
[10] K.S.N .Murthy, A.V.S.Pavan kumar, Gangu Dharmaraju, international journal of engineering, Science
and mathematics, 2017.

[11] Deepiika k.K, Smitha Vinod, Crime analysis in india using data minig techniques , International journal
of Enginnering and technology, 2018.

[12] Hitesh Kumar Reddy ToppyiReddy, Bhavana Saini, Ginika mahajan, Crime Prediction &Monitoring
Framework Based on Spatial Analysis, International Conference on Computational Intelligence Data Science
(ICCIDS 2018).

IJCRT2102057 International Journal of Creative Research Thoughts (IJCRT) www.ijcrt.org 484