2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network
| Format: | General Document |
|---|
| _version_ | 1860797998682865664 |
|---|---|
| building | INTELEK Repository |
| collection | Online Access |
| collectionurl | https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 |
| copyright | Copyright©PWB2025 |
| country | Malaysia |
| date | 2018-03-08 12:22 |
| format | General Document |
| id | 15355 |
| institution | UniSZA |
| originalfilename | THE DEVELOPMENT OF A FAMILY SUPPORT INDEX PREDICTIVE MODEL FOR SUBSTANCE ABUSERS USING MULTIVARIATE ANALYSIS AND AN ARTIFICIAL NEURAL NETWORK |
| person | PDFsam Basic v4.2.10 Siti Nor Fazillah binti Abdullah |
| recordtype | oai_dc |
| resourceurl | https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15355 |
| sourcemedia | Server storage Scanned document |
| spelling | 15355 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15355 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu English UniSZA East Coast Environmental Research Institute application/pdf 1.5 PDFsam Basic v4.2.10 Server storage Scanned document UniSZA Private Access Universiti Sultan Zainal Abidin SAMBox 2.3.4; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) 303 2018-03-08 12:22 THE DEVELOPMENT OF A FAMILY SUPPORT INDEX PREDICTIVE MODEL FOR SUBSTANCE ABUSERS USING MULTIVARIATE ANALYSIS AND AN ARTIFICIAL NEURAL NETWORK 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network Copyright©PWB2025 Multivariate analysis Siti Nor Fazillah binti Abdullah Family Support Index Predictive Model Substance Abusers Multivariate Analysis Artificial Neural Network Substance abuse—Mathematical models Substance abuse—Prediction Family support—Psychological aspects Artificial neural networks—Applications in social sciences Addicts—Rehabilitation—Mathematical models Social networks—Psychological aspects Risk assessment—Mathematical models Mental health—Mathematical models Public health—Statistical methods This study has identified family support as one of the ways to help substance abusers in the rehabilitation process. Nevertheless, not many rehabilitation programs involving family members and no modules and indices of family support for substance abusers have been developed. The objectives of this study are to establish family support data for substance abusers in the study area, to suggest an index of family support, and hence determine the level of family support of substance abusers in the study area and to develop models based on integrated multivariate models and intelligent predictive models for family support of the substance abusers in the study area. This study used collected family support data from 245 respondents who participated at each district in Terengganu from November to December 2016 using a family support instrument that was developed by Farah et al (2016). Several multivariate analyses were performed in this study. These were the factor analysis (AFA) which was used to identify the sources of major factors contributing to family support in the study area (and thus develop family support indices according to each form of family support), and the discriminant analysis (DA) which was used to identify the most significant questions (variables) which have a high variation and prominent role in the determination of the level of family support. The distribution of the family support index was then visualised through the GIS tool. The artificial neural network (ANN) technique was applied to develop several family support-ANN (FS-ANN) prediction models which used scale categories generated from AFA as the output target and with the best performance model developed being chosen as the best prediction model. The results showed that there were three factors of emotional support and instrumental support, two factors of information support and spiritual support, and four factors of support barrier which had major contributions to each form of family support respectively. According to the resulting index, the categories of scale were determined based on five levels which are very good, good, fair, poor and very poor. DA successfully identified the most significant questions with the accuracy of the classification matrix for emotional support, instrumental support, information support and spiritual support using the backward stepwise modes yielding 98.63% (seven variables), 91.06% (seven variables), 94.89% (eight variables) and 100% (eight variables) respectively while the support barrier yielded 91.91% with 13 variables. The most significant questions for each support were used as inputs to develop the FS-ANN prediction models. The five prediction models of family support have two and three different methods with different inputs. The best prediction models were determined by comparing the highest value of R2 and the lowest value of RMSE among the methods. Hence, it can be concluded that the best prediction models for each form of family support using the input derived from the DA backward stepwise mode was where the value of R2 and RMSE for emotional support, instrumental support, information support, spiritual support and support barrier were 0.99 and 0.00, 0.99 and 0.10, 0.94 and 0.18, 0.99 and 0.00 and 0.91 and 0.18 respectively. The FS-ANN prediction models developed can be used for prediction purposes in the future with a lesser number of questions which are more significant. This research has verified that developing this model could help the related stakeholders to understand the level of family support and plan strategies on how to resolve this problem. This will therefore help them make the best decisions on handling needed family support for substance abusers in order to assist involved families in the rehabilitation process of substance abusers efficiently Dissertations, Academic Thesis |
| spellingShingle | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| state | Terengganu |
| subject | Multivariate analysis Substance abuse—Mathematical models Substance abuse—Prediction Family support—Psychological aspects Artificial neural networks—Applications in social sciences Addicts—Rehabilitation—Mathematical models Social networks—Psychological aspects Risk assessment—Mathematical models Mental health—Mathematical models Public health—Statistical methods Dissertations, Academic |
| summary | This study has identified family support as one of the ways to help substance abusers in the rehabilitation process. Nevertheless, not many rehabilitation programs involving family members and no modules and indices of family support for substance abusers have been developed. The objectives of this study are to establish family support data for substance abusers in the study area, to suggest an index of family support, and hence determine the level of family support of substance abusers in the study area and to develop models based on integrated multivariate models and intelligent predictive models for family support of the substance abusers in the study area. This study used collected family support data from 245 respondents who participated at each district in Terengganu from November to December 2016 using a family support instrument that was developed by Farah et al (2016). Several multivariate analyses were performed in this study. These were the factor analysis (AFA) which was used to identify the sources of major factors contributing to family support in the study area (and thus develop family support indices according to each form of family support), and the discriminant analysis (DA) which was used to identify the most significant questions (variables) which have a high variation and prominent role in the determination of the level of family support. The distribution of the family support index was then visualised through the GIS tool. The artificial neural network (ANN) technique was applied to develop several family support-ANN (FS-ANN) prediction models which used scale categories generated from AFA as the output target and with the best performance model developed being chosen as the best prediction model. The results showed that there were three factors of emotional support and instrumental support, two factors of information support and spiritual support, and four factors of support barrier which had major contributions to each form of family support respectively. According to the resulting index, the categories of scale were determined based on five levels which are very good, good, fair, poor and very poor. DA successfully identified the most significant questions with the accuracy of the classification matrix for emotional support, instrumental support, information support and spiritual support using the backward stepwise modes yielding 98.63% (seven variables), 91.06% (seven variables), 94.89% (eight variables) and 100% (eight variables) respectively while the support barrier yielded 91.91% with 13 variables. The most significant questions for each support were used as inputs to develop the FS-ANN prediction models. The five prediction models of family support have two and three different methods with different inputs. The best prediction models were determined by comparing the highest value of R2 and the lowest value of RMSE among the methods. Hence, it can be concluded that the best prediction models for each form of family support using the input derived from the DA backward stepwise mode was where the value of R2 and RMSE for emotional support, instrumental support, information support, spiritual support and support barrier were 0.99 and 0.00, 0.99 and 0.10, 0.94 and 0.18, 0.99 and 0.00 and 0.91 and 0.18 respectively. The FS-ANN prediction models developed can be used for prediction purposes in the future with a lesser number of questions which are more significant. This research has verified that developing this model could help the related stakeholders to understand the level of family support and plan strategies on how to resolve this problem. This will therefore help them make the best decisions on handling needed family support for substance abusers in order to assist involved families in the rehabilitation process of substance abusers efficiently |
| title | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| title_full | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| title_fullStr | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| title_full_unstemmed | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| title_short | 2018_The Development Of A Family Support Index Predictive Model For Substance Abusers Using Multivariate Analysis and An Artificial Neural Network |
| title_sort | 2018_the development of a family support index predictive model for substance abusers using multivariate analysis and an artificial neural network |