Structural equation modelling – artificial neural network based hybrid approach for assessing quality of university cafeteria services

Abstract

PurposeThis study aims to show the effectiveness and applicability of artificial intelligence applications in the measurement and evaluation of university services. Universities can gain competitive advantage through providing their students with quality services in various aspects, such as bookstores, dormitories, recreation centers as well as cafeterias. Among these facilities, university cafeterias are places where students spend a significant amount of time. Therefore, this study aims to integrate artificial intelligence application in the evaluation of university cafeteria services based on students' perceptions with two-stage structural equation modeling (SEM) and artificial neural network (ANN) approach.Design/methodology/approachAn artificial intelligence based SEM-ANN hybrid approach was used to determine the factors that have significant influence on student satisfaction, sufficiency-of-services and likelihood-of-recommendation. Data were collected from 373 students through a face-to-face questionnaire. Initially, four service quality dimensions were attained through factor analysis. Then, hypotheses, which were determined via literature review, were tested through SEM-ANN hybrid approach.FindingsIncorporating the results of SEM analysis into the ANN technique resulted in superior models with good prediction performance. Based on four ANN models created and ANN sensitivity analyses conducted, significant predictors of satisfaction, sufficiency, reliability and recommendation are determined and ranked.Originality/valuePrior studies have assessed service quality using traditional techniques, whereas, this study integrates artificial intelligence in the assessment of higher-educational institutions' services quality. Also, as a distinction from previous studies, this study ranked importance levels of predictor variables through ANN sensitivity analysis.

Examining the slow acceptance of HR analytics in the Indian engineering and construction industry: a SEM-ANN-based approach

“…The present study builds and evaluates a model to comprehend the factors of BI to use and recommend the implementation of HR analytics in various organizations through a multi-analytic SEM-ANN approach. In addition, incorporating the results of SEM analysis into the ANN technique provided superior models with good prediction performance (Uluskan, 2022). Finally, this two-step SEM-ANN approach allowed for an additional examination of the relative significance of the input factors that serve as predictors of adoption of HR analytics among HR professionals in the engineering and construction sectors.…”

“…Moreover, previous studies have evaluated the HR analytics adoption using techniques such as regression analysis and structural equation modeling (SEM) (Vargas, 2015;Vargas et al, 2018). The current research incorporates a hybrid SEM-ANN (artificial neural network) approach which facilitates a better understanding when compared to the single-step SEM technique resulting in superior models with good prediction performance (Uluskan, 2022). The hybrid approach was used to predict the most influential factor of behavioral intention (BI).…”

“…The hybrid approach was used to predict the most influential factor of behavioral intention (BI). Besides, this research evaluated the normalized importance of each predictor input variable through sensitivity analysis (Uluskan, 2022). Covariance-based SEM examines the complex relations between the constructs (Ma et al, 2022) and determines linear relationships, which oversimplifies a complex topic, such as the acceptance of HR analytics.…”

“…An ANN analysis was applied with a single hidden layer (Leong et al. , 2015; Uluskan, 2022) to predict the relevant factors influencing the adoption of HR analytics. The datasets were split at a ratio of 90:10 for training and testing models (Anouze and Alamro, 2019).…”

“…, 2018). The current research incorporates a hybrid SEM-ANN (artificial neural network) approach which facilitates a better understanding when compared to the single-step SEM technique resulting in superior models with good prediction performance (Uluskan, 2022). The hybrid approach was used to predict the most influential factor of behavioral intention (BI).…”