Use of Artificial Neural Network as a Risk Assessment Tool in Preventing Child Abuse

• NCJ Number: 190273
• Author(s): Iraj Zandi
• Date Published: 2000
• Length: 18 pages
• Sale Source: University of Pennsylvania
• Publication Format: Document
• Publication Type: Test/Measurement; Report (Technical)
• Language: English
• Country: United States of America

Annotation

This paper reports on a preliminary effort to explore the feasibility of using artificial neural network (ANN) technology as a risk assessment tool for alleged cases of child abuse.

Abstract

The Third National Incidence Study of Child Abuse and Neglect (NIS-3) -- a congressionally mandated, periodic effort of the National Center on Child Abuse and Neglect -- provides a rich set of data on child abuse and makes the use of ANN a distinct possibility. This data were used in the current research to explore the utility of using ANN as a risk assessment tool. The current study constructed and experimented with several different ANN designs under a variety of conditions. These included several designs of multi-layers neural network (MLNN) by using several different training algorithms and a radial basis network. The data were divided randomly into three groups: A Training Set, a Validation Set, and a Test Set. The procedure was to use the Training Set to train the network and test its veracity by using the Validation Set and the Test Set. The research found that a 31-25-1 MLNN architect, after being trained, was capable of classifying the Training Set 100 percent accurately; however, its performance deteriorated for the Validation Set and the Test Set. Despite this deterioration, the results were very encouraging. The network of Experiment VII-1 was capable of classifying correctly 90 percent of abused cases for the Validation Set and 89 percent for the Test Set. This meant it missed 10 percent of abused children for the Validation Set (false negative) and 11 percent for the Test Set. In addition, it also misclassified 13 percent (false alarm) of those children that the survey did not find to have been abused in the Validation Set as being abused (false negative) and 11 percent of the Test Set. In addition, it also misclassified 13 percent (false alarm) of those children that the survey did not find to have been abused in the Validation Set as being abused (false positive). Although definitive data for the performance of currently used risk assessment tools is unavailable, experts indicate the performance of neural network is a clear improvement over the current practice; however, much research remains to be done. 14 figures, 2 tables, and 4 references