International Research Journal of Engineering and Technology (IRJET) Volume: 07 Issue: 04 | Apr 2020
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
CRIME ANALYSIS & PREDICTION USING DATA MINING Pintu Prasad1, Raj Singh2, Rishabh Singh3, Prof. Reena Kothari4 1Pintu Prasad, Information Technology, Shree L.R. Tiwari College of Engineering 2Raj Singh, Information Technology, Shree L.R. Tiwari College of Engineering
3Rishabh Singh, Information Technology, Shree L.R. Tiwari College of Engineering
4Prof. Reena Kothari, Information Technology, Shree L.R. Tiwari College of Engineering, Mumbai
------------------------------------------------------------------------------***------------------------------------------------------------------------
Abstract: Crimes are a really common global problem affecting the standard of life and therefore the economic process of a
society. the rise of crimes, enforcement is constant to demand advanced global information systems and new data processing techniques to enhance crime analytics and better protect their communities. Although crimes could occur everywhere, Using the concept of knowledge mining, we will analyze previously unknown, useful information from unstructured data. Predictive technique means, using analytical and predictive techniques, to spot crimes and it's just about effective in doing an equivalent. due to the increased rate over the years, we'll need to handle an enormous amount of crime data stored which might be very difficult to be analyzed manually, and also now a day's, criminals are getting technologically advanced, so there's got to use advance technologies to stay police before them. during this paper, the most focus is on the review of algorithms and predicting future crime with past analysis of knowledge.
Keywords:- logistic regression, preprocessing, data collection, data visualization, analysis and prediction. 1.INTRODUCTION: Data Mining is that the procedure which incorporates evaluating and examining large pre-existing databases to get new information which can be essential to the organization. The extraction of latest information is predicted using the prevailing datasets. Many approaches for analysis and prediction in data processing had been performed. But, many few efforts have made within the criminology field. Many few have taken efforts for comparing the knowledge of these approaches produce. The police headquarters and other similar criminal justice agencies hold many large databases of data which may be wont to predict or analyze the criminal activity involvements within the society. The criminals also can be predicted supported the crime data. This paper presents a survey on Crime analysis and crime prediction using several data processing techniques. Our main contribution during this paper is to propose an approach supported data processing and classification method which becomes tougher when handling the large amount of knowledge, thus machine learning reduces the quantity of your time to predict crime and analysis and plot the graph of crime year wise. The organization of the paper is as follows. Section II consists of some important data processing techniques. Section III consists of an outline of the Knowledge Discovery process. The logistic regression Methods are discussed in section IV. The methods applied in crime domain are discussed in section V and therefore the paper is concluded
2. EXISTING SYSTEM: 1) Agarwal et al. used the rapid miner tool for analyzing the rates and anticipation of crime rate using different data processing techniques. Their work done is for crime analysis using the K-Means Clustering algorithm. the most objective of their crime analysis work is to extract the crime patterns, predict the crime supported the spatial distribution of existing data and detection of crime. Their analysis includes the tracking homicide crime rates from one year to subsequent. 2) Satyadevan et al. has done a piece which can display high probability for crime occurrence and may visualize crime-prone areas. rather than just that specialize in the crime occurrences, they're focusing mainly on the crime factors of every day. They used the Naïve Bayes, Logistic Regression and SVM classifiers for classification of crime patterns and crime factors of every day. Their method consists of a pattern identification phase which may identify the trends and patterns in crime using the Apriori Algorithm. The prediction of crime spots is completed with the assistance of the choice Tree algorithm which can detect the crime possible areas and their patterns. © 2020, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 5732
International Research Journal of Engineering and Technology (IRJET) Volume: 07 Issue: 04 | Apr 2020
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
3. METHODOLOGY: Crime analysis and prediction using data processing system have mainly five steps. the primary step is that the data collection step which we've taken from the Chicago data portal. The second step is that the data smoothing step. The third step is that the prediction step, we've used logistic regression step. The fourth step is data visualization.
3.1DATA COLLECTION: The data utilized in this research comes from the Chicago data portal and website for open data about crime and policing in United America. This dataset reflects reported incidents of crime (except for murder where data exists for every victim) that occurred within the City Of Chicago from 2001 to present, minus the foremost recent days. Data is extracted from the Chicago local department portal. .
Fig 1 :Data From Portel
3.2.SMOOTHING : Smoothing may be a technique that's wont to eliminate noise from a dataset. There are many algorithm and methods to accomplish this but all have an equivalent general purpose of roughing out the sides or 'smoothing' some data. there's reason to smooth data if there's little to no small-scale structure within the data. the info danger to the present thinking is that one may skew the representation of the info enough to vary its perceived meaning, so for the sake of scientific honesty its is an important to at the very minimum explain one’s reason’s for employing a smoothing algorithm to their dataset. So for the smoothing purpose, we are using Dropna Method which available in pandas package. Pandas is one among those packages and makes importing and analyzing the info easier.
Fig 2:: Smooth Data From Dropna Method
3.3 PREDICATION: By using the various parameter for prediction we checked various algorithm like Logistic Regression Model, Decision Tree and Random Forest. Regression technique are often adapted for predication. But Logestic Regression show more accuracy on Chicago Dataset as compare to other two algorithm. Logestic analysis are used for prediction whether future crime increase or decrease supported current bookcase. In data processing independent variables attributed already known and response © 2020, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 5733
International Research Journal of Engineering and Technology (IRJET) Volume: 07 Issue: 04 | Apr 2020
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
variables attributed what we would like to predict.Once Accuracy and Predicton is completed we've created a model by using Joblib is the quality way of serializing objects in Python. you'll use the Joblib operation to serialize your machine learning algorithms and save the serialized format to a file. Later you'll load this file to deserialize your model and use it to form new predictions.The same model type often is employed for both regression and classification. This method is predicated on the booked case entered by the user. After entering the amount they use the various parameter for prediction like Not Arrested, Arrested and Efficiency as shown in below
Fig 3: Different Model Accuracy
Fig 4: Predication Process And Predication GUI
3.4 DATA VISUALIZATION: Data Visualization is important for representing insight from data during a graphical manner. With an outsized amount of knowledge within the dataset, one among the best challenges is to simply communicate the hidden pattern and findings easily and understandably.to visualize the info there are many visualization techniques available.
Fig5: Logistic Regression Graph And Bar Graph © 2020, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 5734
International Research Journal of Engineering and Technology (IRJET) Volume: 07 Issue: 04 | Apr 2020
www.irjet.net
e-ISSN: 2395-0056 p-ISSN: 2395-0072
Fig6: Correlation Matrix And Basic Plotting Graph
Conclusion: In today’s world, where the typical quantity of knowledge that an individual handles has been increasing by leaps and bounds over the past few years, the use of knowledge mining techniques to extract useful information from the large amounts of data becomes important. This project mines the large amounts of data by first generating it within the sort of a dataset then preprocessing it. the varied data processing techniques, algorithms and models mentioned when applied on such datasets produce results which might be of great potential use to enforcement agencies especially. last, thus, we hope that this project performs its functions well, and works with the very best efficiency possible which it surely proves to be a boon to enforcement agencies. The functionalities of this project are often scaled up within the future. These functionalities might be Real-time data analysis of crime data: this might help us obtain crime patterns and forecasts of the longer term instantly using real-time datasets. data processing of social media to get datasets, then preprocess and analyse them to identify trends of the present crime situation during a particular place or region. Compare and display the results of all available and applicable forecasting, predicting and classification models side by side, such the user can select any of these methods.
Future work: The proposed project mainly focuses on highlighting the crime rates around a specific state .So as per the longer term vision, another a part of the country and world are often taken into consideration. immediately it's specifically sure to a couple of crimes. Other crimes are often included further. There are certain crimes andcases which are unheard and unregistered round the globe and if they're taken into attention theaccuracy of the crime rates are often improved. Currently, the dataset consists of only registred crimes but this will be expanded to incorporate more crimes within the future. and that we will attempt to expand our project as criminal profilling
References: [1] Alkesh Bharati1, Dr Sarvanaguru RA.K2, “Crime Prediction and Analysis Using Machine Learning”. (IRJET) e-ISSN: 23950056 | Sep 2018| (IRJET) [2] ChhayaChauhan1, SmritiSehgal2: A REVIEW: Crime Analysis Using Data Processing Techniques And Algorithms. ICCCA2017) [3] Gaurav Govindaswamy1, Santhosh Kumar P3” A Survey on Crime Data Analysis Using data processing Techniques”. Int. Journal of Engineering Research and Application) August 2017, pp.30-34. 4] H. Benjamin Fredrick David1 and A. Suruliandi2, “Survey On Crime Analysis And Prediction Using Data Processing Techniques”. ICTACT JOURNAL ON SOFT COMPUTING, APRIL 2017, [5] Sunil Yadav, Meet Timbadia,.” Crime Pattern Detection, Analysis & Prediction” ICECA 2017
© 2020, IRJET
|
Impact Factor value: 7.529
|
ISO 9001:2008 Certified Journal
|
Page 5735