International Journal of Innovative Technology and Interdisciplinary Sciences

Regression-Based Machine Learning for Predicting Prior Convictions from Administrative Criminal Justice Records

Thu, 02 Apr 2026 00:00:00 +0200

Predictive analytics is emerging in current police and court proceedings where data is used to analyse crime trends and court results; predictive, regression-based machine learning models use these data to predict dependent variables. This study explores the predictive use of regression-based machine learning models to predict counts of prior convictions based on structured data within a criminal justice dataset. The analysis is done on the Criminal Justice Dataset that is a publicly available data set with roughly 200,000 court cases data with demographic characteristics, criminal history factors, offense parameters and court decisions variables. Using descriptive statistical analysis estimates, correlation univariate tests and multiple regression predictive models the relationships are explored between the independent predictors and the dependent counts of prior convictions. Three model builds are used; the first is an OLS model to provide a comparison baseline followed by the use of a Ridge regression model with L2 regularization penalty and an ElasticNet model with L1 and L2 as a way of deploying multiple robust models to produce the best fit and prediction. We evaluate the performance of the model using standard regression metrics: coefficient of determination (R²), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE).

Comparative Seismic Performance and Retrofit Guidelines for URM, CM, and RC Buildings Based on 24 Real Post-Earthquake Case Studies

Marjo Hysenlliu, Enio Deneko, Altin Bidaj, Huseyin Bilgin — Wed, 08 Apr 2026 00:00:00 +0200

The 2019 Mw 6.4 Albania earthquake revealed considerable vulnerability encompassing a range of different structural typologies, predominantly among buildings assigned as Damage State 4 (DS-4). Numerous efforts investigate common construction typologies such as: unreinforced masonry (URM), confined masonry (CM), and reinforced concrete (RC), however, there is a gap in the ability to comparatively quantify residual seismic and retrofit response behaviours from these typologies in a common analytical framework. A comparative evaluation of 24 actual post earthquake buildings (8 URM, 8 CM, 8 RC) has been performed uniformly applying nonlinear static (pushover) procedure in the scope of Eurocode 8 – Part 3. Normalized base shear capacity (Vb/W), inter-story drifts at Damage Limitation, Significant Damage and Near Collapse limits and global ductility (μ) were obtained for each building by bilinear idealizations of capacity curves. Statistical descriptors were also calculated for each typology (mean, standard deviation, coefficient of variation), and hypothesis testing comparisons were conducted to assess if (i) residual strength capacity (at DS-4) is hierarchically ordered as RC > CM > URM, and (ii) increasing levels of ductility after retrofit are statistically significantly different among typologies. The results support a statistically significant trend in residual strength and ductility capacity, with RC systems showing a higher normalized base shear, and URM buildings receiving the greatest percentage increase in ductility after retrofit. To express retrofit performance in a similar quantitative way, a single performance indicator called the Retrofit Efficiency Index (REI) is developed that aggregates the strength and ductility ratios within one dimension. Such a performance indicator emphasizes the advantage of the typology and also establishes a useful decision making index for heterogeneous stock of buildings. The results encapsulate real post earthquake evidence into statistically, supported performance ranges, and summarises comparative guidelines. Instead of proposing a new modelling methodology, the new corpus of results provides a rigorous quantitative framework based on uniformly analysed real damaged buildings, and provides benchmark standards directly relevant for engineering evaluation and retrofit ordering in seismic regions.

Urban Change and Social Cohesion in Post-Socialist Contexts: An Empirical and Theoretical Assessment

Otjela Lubonja, Sabina Bollano, Dora Foti, Ginevra Amendolagine — Thu, 09 Apr 2026 00:00:00 +0200

This paper examines an underexplored dimension of the social impacts of gentrification, namely the neighbourhood social cohesion, in Tirana, the post-socialist capital of Albania, which experienced a rapid development during the last decade. Based on a survey with 201 residents of neighbourhoods with distinct levels of urban transformation, it explores the correlation between gentrification pressures, such as increasing housing prices, commercialization, tourism pressure and intensification of redevelopment and social cohesion. The latter is measured with the help of a Social Cohesion Index (SCI) relying on four dimensions: trust, belonging, interaction and attachment to place. The regression and exploratory mediation analyses are applied to assess statistical associations and indirect pathways while also exploring the intermediary effect of cultural displacement. We found that housing affordability is the strongest predictor of declines in social cohesion, and that cultural displacement appears to play a mediating role in the relationship of urban redevelopment intensity to social cohesion. Socio-economic groups show different perceptions of urban change. Long-term and lower-income residents are more likely to perceive urban change negatively and higher-income residents tend to regard urban renewal more favourably. This paper contributes to the growing literature on post-socialist cities by developing a context-sensitive analytical framework for understanding gentrification in rapidly transforming urban contexts. More generally, it highlights how perception-based mechanisms connect macro-level urban transformation with meso-level neighbourhood dynamics and micro-level social outcomes. It therefore helps to reconceptualize gentrification as a condensed and perception-mediated process.

Systematic Review and Meta-Analysis of Pharmacologic Cardioprotection in Breast Cancer Therapy: Evidence from Real-World Data

Benard Shehu, Klerida Shehu, Fatjona Kraja, Bledar Kraja — Sun, 12 Apr 2026 00:00:00 +0200

Pharmacologic cardioprotection during anthracycline and/or anti‑HER2 therapy for breast cancer remains incompletely characterized, particularly in real‑world settings. This study synthesizes trial evidence and validates it against an Albanian cohort to define the potential benefit and residual burden of cardiotoxicity. A random‑effects meta‑analysis of three comparative trials demonstrated that prophylactic cardioprotective medications were associated with a significant attenuation of left ventricular ejection fraction (LVEF) decline (pooled mean difference = 2.45 percentage points, 95% CI 1.56–3.34; I² = 21.8%). In a parallel retrospective analysis of 314 Albanian breast cancer patients, symptomatic cardiotoxicity occurred in 16.6%, and a clinically relevant ≥10‑point LVEF decline within 12 months was observed in 12.1%. These findings confirm a modest but consistent cardioprotective effect from prophylactic interventions in clinical trials, yet the substantial real‑world burden affecting approximately one in six patients symptomatically and one in eight with significant LVEF decline highlights a critical gap between trial efficacy and routine practice, underscoring the need for systematic implementation of cardioprotective strategies.

GAN-Augmented Attention-Based CNN for Non-IID Federated Diabetic Retinopathy Classification

Aishwarya Mane, Swati Shekapure — Sun, 19 Apr 2026 00:00:00 +0200

Scalable and privacy-preserving diagnostic models are essential for diabetic retinopathy "DR" screening across multi-centre healthcare institutions, where data are heterogeneous and non-identically distributed (non-IID). Although centralised deep learning methods achieve high accuracy, they are impractical due to privacy constraints and cross-institutional data variability. To address this, we propose an attention-enhanced EfficientNet-B0 integrated with Federated Proximal Optimisation (FedProx) and GAN-based minority augmentation under simulated label-skew non-IID conditions. Experiments on APTOS 2019 and IDRiD datasets, partitioned into four federated clients, achieved 77.6% accuracy and 0.91 macro-AUC. The framework improved minority recall by 5–8% and reduced convergence variance by 44%, demonstrating stable and practical federated DR grading for real-world clinical deployment. The model was trained over 50 communication rounds across four simulated federated clients under controlled label-skew partitioning. Compared to FedAvg under identical non-IID settings (72.1% accuracy, 0.85 macro-AUC), the proposed framework demonstrates statistically significant improvement (p < 0.05).

Modeling and Prediction of Student Achievement: The Role of Self-Efficacy, Motivation, Time Management, and Engagement

Anisa Subashi, Edit Bregu, Nazmi Xhomara — Sun, 19 Apr 2026 00:00:00 +0200

This study investigates the role of time management, self-efficacy, motivation, and engagement in predicting student achievement. Data were collected through an online questionnaire administered to bachelor’s students in Business Informatics, Economics, and Law at a large higher education institution. The results indicate that students scored above average across all variables. Among the examined factors, self-efficacy and motivation emerged as stronger determinants of student achievement compared to time management and engagement. In addition, self-efficacy and motivation showed stronger positive correlations with academic performance and demonstrated greater predictive power. These findings contribute to the existing literature on academic performance by highlighting the relative importance of psychological and behavioral factors. The study provides practical insights for educators and institutions, emphasizing the need to support students’ motivation and self-efficacy in order to enhance academic outcomes.

Urban Plurality Without Polycentricity: A Cross-Dimensional Analysis of Albania’s Deviation from the Polycentric Urban Region Model

Gentian Kaprata — Fri, 24 Apr 2026 00:00:00 +0200

The topic of polycentricity remains dominant in European urban planning discussions; however, previous studies have focused only on larger, wealthier, and administratively divided metropolitan areas. There has been little research conducted to date exploring the dynamics of the failed emergence of polycentricity in small-scale, centralized, and newly urbanized countries. This article will fill the void by analyzing the Albanian case. Rather than applying the European approach to testing, the Polycentric Urban Region concept will serve as a theoretical framework for analysis across six aspects: plurality of centers, complementarity, inter-urban relationships, hierarchy, coordination, and matching scales. Based on three propositions, qualitative analysis will be undertaken through three periods of Albanian politics. According to the findings, morphological plurality can be observed in Albania, but not functional polycentricity, as metropolitan primacy, lack of differentiation, radial connections, fragmented governance, and scalar mismatches maintain the stable monocentric order. Two terms will be introduced as outcomes of the study: structural scalar lock-in and pseudo-polycentric equilibrium. Comparative assessment with Kosovo, Montenegro, North Macedonia, and Serbia supports the hypothesis.

Intelligent Solar Power Forecasting Using a Neuro-Evolutionary DE-Optimized TCN–LSTM Hybrid Model

Sun, 26 Apr 2026 00:00:00 +0200

Solar photovoltaic (PV) generation changes a lot, which makes it hard to keep the grid stable, trade energy, and manage storage. This shows how important it is to be able to accurately predict the short term. This research presents DE-TCN-LSTM, a hybrid deep learning framework that integrates Temporal Convolutional Networks (TCN), bidirectional Long Short-Term Memory (LSTM), and Differential Evolution (DE) for the automation of hyperparameter optimization. Rather than introducing novel model architectures, this study presents a regulated comparison of Differential Evolution (DE) with Particle Swarm Optimization (PSO) and Genetic Algorithms (GA) for the optimization of Temporal Convolutional Networks-Long Short-Term Memory (TCN-LSTM) models in medium-scale photovoltaic output forecasting. The experimental evaluation uses one year of operational data from a 7.7 MW solar plant. A 6-fold rolling-origin cross-validation with 10 repetitions yields 60 runs per model. The DE-TCN-LSTM model achieves RMSE = 59.2 ± 3.0 kWh, MAE = 31.7 ± 1.8 kWh, and R² = 0.981 ± 0.003, with a normalized nMAE of 0.41%. Statistical analyses, encompassing ANOVA, Wilcoxon tests with Holm correction, Diebold-Mariano, and bootstrap confidence intervals, validate substantial enhancements compared to all baseline methodologies. RMSE is reduced by 32.1% compared to TCN-LSTM, 13.8% compared to PSO-TCN-LSTM, and 21.4% compared to GA-TCN-LSTM. Cross-dataset validation using SURFRAD and BSRN datasets shows robust generalization, with RMSE degradation limited to 5.4% or less across different climatic conditions. These findings highlight the importance of optimizer selection in improving deep hybrid models for large-scale PV forecasting.