International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 09 Issue: 12 | Dec 2022
p-ISSN: 2395-0072
www.irjet.net
Review of Algorithms for Crime Analysis & Prediction Siddik Aiyub Patel1, Jayanth R2, Pawan T3, Dr. E. Praynlin4 123Students,
Computer Science and Engineering, T. John Institute of Technology, Bengaluru, Karnataka, India Professor, Computer Science and Engineering, T. John Institute of Technology, Bengaluru, Karnataka, India ---------------------------------------------------------------------***--------------------------------------------------------------------4Asst.
Abstract - Crime prediction is a rigorous approach to
for further processes. This is known as "data preparation". Following that, modelling strategies must be chosen, at which point several tools and techniques that fall within the machine learning algorithm are introduced. Data is assessed once data modelling is applied, and then it can be used for related purpose. Machine learning uses a variety of algorithms and methods that are applied to data to gain knowledge and experience. The two fundamental types of machine learning algorithms are supervised learning and unsupervised learning. In supervised learning, an input is mapped to an output after the data is taught with correct and incorrect responses. Unsupervised learning, on the other hand, requires independent learning because there is no prior knowledge.
spotting patterns and trends in crime. This paper discusses various technologies that can be utilized to create a system for crime prediction. By constructing a Crime Prediction System, it accelerates the investigation of crimes and lowers the crime rate. We use a variety of methodologies, each of which is based on previously reported and recorded data containing time and place. The Crime Prediction System collects recorded data and analyses it using a variety of methodologies before forecasting the patterns and trends of crime using any of the strategies explored below.
Key Words: Crime prediction, Clustering, Crime rate, Data Mining, Machine Learning, Deep Learning
Data mining for crime prediction includes a number of techniques. Here, we have covered some of the commonly used techniques.
1. INTRODUCTION When crimes are committed regularly in a society, they in certain ways have an impact on organizations and institutions. As a result, it's important to research the connections and contributing elements to various crimes in order to effectively forecast and prevent them. The approach to predictive policing that law enforcement agencies are taking recently is becoming more pragmatic and data-driven. The foundational work of crime analysts, however, continues to be challenging and more often labouring. The objective of this paper is to present a thorough examination of theory and research about the prevention of crime in society and to put into practice various data analysis algorithms that address the relationships between crime and the patterns that crime follows. We have discussed various data mining algorithms along with deep learning technique to widen the scope of review
2.1 CLASSIFICATION A classification approach represents and discerns data classes or ideas. Input data are classified into groups. Data set is divided into two types: Training data set and Testing data set. Former is used to train the model and the latter is used to test model and check its accuracy. Various approaches to this process are: 1) Decision Tree: A decision tree is a type of tree structure that resembles a flowchart, where each internal node represents a test on an attribute, each branch a test result, and each leaf node (terminal node) a class label or result. Data with few dozen to many thousands of dimensions can be handled by decision trees. To classify an instance, one tests the attribute given by the root node of the tree before continuing down the branch of the tree that corresponds to the attribute's value. The subtree residing at the new node is then subjected to the same procedure once more.
2. DATA MINING METHODS The process of retrieving information from a database or any other piece of data using understanding to produce a number of outputs is known as data mining. It is nothing more than a process of gathering information and identifying patterns in the raw data in order to come to or make an effective judgement. Some fundamental steps involved are: Understanding your work target is the first step. Create a quick plan and decide on data mining objectives. In second step, the emphasis is placed on data collecting, data visualization and data examination. After reviewing the data, at third step, the data scientist starts by examining and data cleaning, keeping only the necessary data that may be used
© 2022, IRJET
|
Impact Factor value: 7.529
Yehya analysed and predicted San Francisco crime data using characteristics including longitude (X), latitude (Y), address, day of the week, date (YYYY-mm-dd: hh: MM: ss), district, resolution, and category. To classify the accuracy and prevent overfitting, the study used a variety of approaches, including principal component analysis. Additionally, he applied four alternative classifiers—K-NN, XGB Decision Tree, Bayesian, and Random Forest—to the task, with the Random Forest classifier yielding the best
|
ISO 9001:2008 Certified Journal
|
Page 284