Goodness-of-fit tests are commonly used to assess the adequacy of multinomial logistic regression models. This study proposes a hybrid > goodness-of-fit test that combines covariate-space partitioning with grouping based on predicted probabilities. However, the adherence of the proposed hybrid test statistic to its hypothesized null distribution has not yet been established. Thus, this study evaluates its adherence to the hypothesized null distribution for three different numbers of groups, \(G=8, G=10\), and \(G=12\). A simulation study was conducted using Ward's hierarchical clustering technique with the Euclidean distance measure, sample sizes of 100 and 400, and various covariate distributions, including uniform, normal, and chi-square distributions. The test was evaluated based on empirical rejection rates and sample means compared with the theoretical means of the corresponding chi-square distributions, given by \((G-2)(c-1)\). The simulation results indicate that the three choices of group numbers ( \(G=8,10\), and 12) show similar patterns, with rejection rates generally close to the nominal 5% significance level and mean teststatistic values generally close to their theoretical expectations. These findings suggest that the choice among \(G=8, G=10\), and \(G=12\) has little effect on the null distribution of the hybrid test statistic.
Principal component analysis (PCA) is a widely used multivariate statistical technique for reducing the dimensionality of large datasets while retaining most of the original information. In this study, PCA was applied to identify the underlying socioeconomic components of household characteristics using data from 368 randomly selected households in Villanueva, Misamis Oriental, drawn from a population of 8,293 households. The suitability of the dataset was assessed using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Re- Bartlett's test of sphericity prior to analysis. The overall KMO value of 0.827 and a significant Bartlett's test ( \(\chi^{2}=1457, p
Against the backdrop of China's dual-carbon goals, climate policy uncertainty represents a critical external constraint on corporate green transformation, yet firms differ considerably in their perceptions of this uncertainty. Drawing on upper echelons theory and resource dependence theory, this study investigates how top management team (TMT) background characteristics influence corporate green investment. Using a sample of Chinese A-share listed firms from 2004 to 2023, we examine the mediating mechanisms and boundary conditions of this relationship. The results reveal three key findings. First, TMT functional, financial, academic, and overseas backgrounds are associated with lower perceived climate policy uncertainty, with managerial power amplifying the effects of the first three backgrounds but dampening that of overseas background. Second, perceived uncertainty mediates these relationships, yet the mediation pattern diverges according to executives' strategic orientations. Executives with functional and financial backgrounds exhibit an inconsistent mediation pattern (negative total effect), whereas those with academic and overseas backgrounds display a consistent mediation pattern (positive total effect). Third, industry carbon intensity moderates these mediated relationships by attenuating or enhancing the effects of different backgrounds and weakening the direct inhibitory effect of perceived uncertainty on green investment. By uncovering the micro-level mechanisms through which firms respond to climate policy, this study provides novel empirical evidence and policy implications for understanding corporate green transformation.
Generation Alpha is approaching higher education after growing up in a learning ecology shaped by mobile technologies, platformized communication, multimodal content, and artificial intelligence. Existing learning readiness and digital readiness models were developed mainly for earlier cohorts and may underrepresent AImediated learning, multimodal academic participation, digital selfregulation, and social-emotional readiness. This paper proposes a systematic and statistically oriented framework for constructing a Generation Alpha Higher Education Readiness Model (GA-HERM) and a composite Generation Alpha Readiness Index (GARI). The proposed design follows PRISMA 2020 and integrates systematic literature review, bibliometric mapping, qualitative content analysis, meta-synthesis, and weight analysis/meta-analysis. Its core contribution is an auditable evidence-to-construct-to-weight pathway in which reported Generation Alpha characteristics are extracted, coded, clustered into constructs, translated into indicators, and weighted using pooled standardized evidence where available. The expected output is a formative readiness model in which digital fluency, AI collaboration readiness, multimodal learning orientation, cognitive adaptability, digital self-regulation, and social-emotional/ethical readiness are treated as evidence-derived components. The novelty lies in shifting readiness measurement from top-down adaptation of prior models to bottom-up, cohort-sensitive, and evidence-weighted design. This paper is presented as a methodological protocol and statistical framework proposal: the systematic review has not yet been completed, all findings are prospective, and the reported formula and illustrative results are theoretical specifications pending empirical execution of the proposed pathway.
Accurate electricity consumption forecasting is essential for effective energy management, power system planning, and sustainable economic development. Hence, this study aims to forecast electricity consumption using two hybrid forecasting models, namely, EMD-ARIMA and EMD-GMDH. Monthly electricity consumption data from Peninsular Malaysia (January 2011-December 2022) and Thailand (January 2002-December 2021) were used to evaluate the forecasting performance of the proposed models in comparison with the standalone ARIMA and GMDH models. The forecasting performance was assessed using the RMSE, MAE and MAPE. The results indicate that the hybrid EMD-GMDH model outperformed the other models by producing lowest forecasting errors, demonstrating its effectiveness for electricity consumption forecasting.
