Risk and Protective Factors of Child Delinquency

Abstract

The Rapid growth of Juvenile delinquency or criminal offense by minor children below 18 years age is widely noticed in India. There are many factors like family & parents, education levels, financial problems, peer groups, mental issues, drugs, media, internet, easy access to online videos etc contributed for the immoral ideas of the teens. In this research paper, we make analysis of two major risk factors, family background & education levels of juveniles involved in crimes. We discover strong rules in Juvenile delinquency dataset of India by Association rule mining, a rule-based machine learning method.

Keywords: Juvenile delinquency, Data mining, machine learning, Association rule mining, Support, Confidence, WEKA tool.

Introduction

The juvenile crime rate in India is risen by over 50% in the last five years. Delhi Nirbhaya's case and the Ryan International School case are few of the heinous crimes that remained in minds of people and media for many years. Shockingly in these abhorrent crimes children below 18 years age are involved.

Get quality help now

Writer Lyla

Verified writer

Proficient in: Child

5 (876)

“ Have been using her for a while and please believe when I tell you, she never fail. Thanks Writer Lyla you are indeed awesome ”

+84 relevant experts are online

Hire writer

The child’s smooth and soft minds can simply be carve and subjected to turn towards criminal offenses.

The criminal idea of the youth is contributed by diverse risk factors like family & parents, education, financial problems, peer groups, mental issues, drugs, media, etc. [1]. Children learn good or bad characteristics from the family. The rejected children of parents or living with guardians or homeless are found at high risk of becoming delinquent. In building the youth, education also plays a vital role. Some key elements in the education system like strict suspension, academic failure, disciplinary methods, and school dropout also add juvenile delinquency cases.

As per National Crime Records Bureau NCRB, India [2] over 31396 juvenile cases are registered in 2015. Fig 1 presents major juveniles crime records for the years 2010-15 in India.

Related Work

Home and family are the primary growth centers of children. The overall development of children is done by parents and family. They protect children from negative criminal activities. If the family does not provide the necessary guidance and support, the children tends towards criminal actions family risk factors that motivate juvenile crimes are parents monitoring level, discipline of child by parents, particularly crude punishments, divorce cases of parents, parent clashes, criminal parents or siblings, etc. [3]. School education is an important element in children's life. In the school-to-prison pipeline, the important components that have been found are strict suspension & disciplinary methods, academic failure and school dropout. Script et al [4] reveal, if children drop school, then they are eight times more liable to be offended than children passing high school. Several data mining techniques were used for crime detection and analysis [5]. In past only few researchers investigate the crime by using association rule mining technique of data mining. In this paper we make analysis of two major risk factors, family background & education levels of juveniles involved in crimes to discover strong rules in Juvenile delinquency dataset of India.

Methodology

Data mining process is the extraction of valuable knowledge from large databases and data warehouses. Few functionalities of data mining used to extract hidden knowledge are generalization, association, classification, clustering etc. which use different algorithms and methods of machine learning. Data mining is one of the useful tools in different fields such as banking, insurance, marketing and manufacturing industry, sales, CRM, crime analysis and fraud identification, transportation, health care, bioinformatics [6]. Association rule mining [7] is a rule-based machine learning method for finding associations and relations, frequent patterns among item sets in large databases using some measures like support, confidence. Association rules are if/then statements that have two parts, an antecedent (if) and a consequent (then). The rule form: Antecedent → Consequent [support, confidence]An example: buys(x, “pencil”) ^ buys(x, “eraser”) → buys(x, “sharpener”) [60%, 75%]The juvenile dataset is obtained from the crime database of India [2]. Preprocessing steps are applied on juvenile dataset in WEKA tool. The clean juvenile data is used for association task to produce association rules meeting minimum support (65%) and confidence (95%) specified by user. The obtained rules are investigated and appraisal is made for making inferences and interconnections between juvenile crimes and two major risk factors, family background and education levels.

Fig 2 shows an overview of association rule mining of juvenile dataset.WEKA a data mining tool is made with java programming language. The full form of WEKA is Waikato Environment for Knowledge Analysis. The University of Waikato in New Zealand creates the WEKA software. WEKA implements many algorithms from machine learning for data preprocessing, association, classification, clustering etc. WEKA also contains visualization functions to see huge data [7]

Experiments, Results and Discussions

The crime database of India is downloaded from the NCRB and government ministry website [2]. The crime data contains many crime statistics for the years 2001-15. Juvenile crime records (527) are gathered by filtering the crime database and saved in excel sheet. Fig 3 shows Indian juvenile crime data (part) during the years 2001-15. The juvenile excel file is saved in CSV format for further processing in WEKA tool. During preprocessing steps, we delete unnecessary attributes (area, year, subgroup) and discretize some attributes (education level attributes illiterate, upto primary, above primary but below matric, matric or above, family background attributes homeless, living with parents living with guardian and total crime) and prepare the juvenile data for association rule mining task. The CSV file after preprocessing is saved as ARFF file. Association rules mining task is applied on the juvenile ARFF file with selected user parameters min support (65%), min confidence (95%) and the top five rules are generated. Fig 4 presents the WEKA output file with association rules relating the juvenile family background & education levels and total crimes. The top five rules generated with juvenile family background and education levels combined with crimes discovers more illiterate juvenile’s are in the first category to make crimes. The children who live guardians are in second group youth incline to crimes. The children who study upto primary school are in third tier youth to offend juvenile crimes. Similarly on investigating the fourth and fifth top rules we derive youth who finish metric and above education are in fourth level and juveniles who lives with parents are in fifth zone are making juvenile crimes. The outcome association rules are similar in lines of the overall crime statistics of India given by NCRB [2] and Ministry of statistics. The reports of crime analysis of data by NCRB [1], India regarding the family background of juveniles arrested in 2015 shows that 86% of the children apprehended lived with their parents and one-third of total crimes offend by minors study upto primary school.

Conclusion and Future Work

We analyze and examine the association of family background & education level risk factors that cause juvenile delinquency in India. The experimental results of the study show a strong association of family background & education levels and juvenile delinquency in India. The association rules generated are similar in lines with India crime statistics given by the Ministry of statistics and NCRB. In the future, similar studies can be taken for analyzing many risk factors with different techniques of data mining to examine the link between risk factors and juvenile delinquency which will help in taking early prevention steps to mitigate juvenile crimes