Methodology For Comparative Evaluation Research Articles

Adaptive machine learning for forecasting in wind energy: A dynamic, multi-algorithmic approach for short and long-term predictions

This study elucidates the formulation and validation of a dynamic hybrid model for wind energy forecasting, with a particular emphasis on its capability for both short-term and long-term predictive accuracy. The model is predicated on the assimilation of time-series data from past wind energy generation and employs a triad of machine learning algorithms: Artificial Neural Network (ANN), Support Vector Machine (SVM), and K-Nearest Neighbors (K-NN). Empirical data, harvested from a 2 MW grid-connected wind turbine, served as the basis for the training and validation phases. A comparative evaluation methodology was devised to scrutinize the performance of each constituent algorithm across a diverse array of metrics. This evaluation framework facilitated the identification of individual algorithmic limitations, which were subsequently mitigated through the implementation of a dynamic switching mechanism within the hybrid model. This innovative feature enables the model to adaptively select the most efficacious forecasting technique based on historical performance data. The hybrid model demonstrated superior forecasting accuracy in both, short-term energy forecasts at 15-min intervals over a day, and in broad, long-term. It recorded a Normalized Mean Absolute Error (NMAE) of 5.54 %, which is notably lower than the NMAE range of 5.65 %–9.22 % observed in other tested models, and significantly better than the average NMAE found in the literature, which spans from 6.73 % to 10.07 %. Such versatility renders it invaluable for grid operators and wind farm management, aiding in both operational and strategic planning. The study's findings not only contribute to the existing body of knowledge in renewable energy forecasting but also suggest the hybrid model's broader applicability in various other predictive analytics domains.

Comparative Evaluation of VAEs, VAE-GANs and AAEs for Anomaly Detection in Network Intrusion Data

With cyberattacks growing in frequency and sophistication, effective anomaly detection is critical for securing networks and systems. This study provides a comparative evaluation of deep generative models for detecting anomalies in network intrusion data. The key objective is to determine the most accurate model architecture. Variational autoencoders (VAEs), VAE-GANs, and adversarial autoencoders (AAEs) are tested on the NSL-KDD dataset containing normal traffic and different attack types. Results show that AAEs significantly outperform VAEs and VAE-GANs, achieving AUC scores up to 0.96 and F1 scores of 0.76 on novel attacks. The adversarial regularization of AAEs enables superior generalization capabilities compared to standard VAEs. VAE-GANs exhibit better accuracy than VAEs, demonstrating the benefits of adversarial training. However, VAE-GANs have higher computational requirements. The findings provide strong evidence that AAEs are the most effective deep anomaly detection technique for intrusion detection systems. This study delivers novel insights into optimizing deep learning architectures for cyber defense. The comparative evaluation methodology and results will aid researchers and practitioners in selecting appropriate models for operational network security.

Value of information analysis for assessing risks and benefits of nanotechnology innovation

BackgroundDecisions on adoption of technological innovation are difficult for manufacturers, especially for small and medium enterprises (SMEs) who have limited resources but often drive product development. Decision analytic methods have been applied to regulatory issues in the nanotechnology sector but such applications to market innovation are not found in the literature. Value of information (VoI) is a decision analytic method for quantifying the benefit of acquiring additional information to support such analyses that can be used to help in a wide range of manufacturing decisions.ResultsThis paper develops a VoI methodology for comparative evaluation of technological alternatives and applies it to a real case study aimed at the selection between a coating system containing nano-TiO2 and alternative conventional paints. The aim of this approach is to aid SMEs and larger industries in deciding whether to further develop the nano-enabled product and in evaluating to which extent investing in more research about risks and/or benefits would be worthwhile.ConclusionsResults demonstrated how prioritization in information gaining can improve risk–benefit analyses and impact on both risk management and innovation decision making. By applying the proposed methodology, SMEs and larger industries might easily identify optimal data gathering and/or research strategies to formulate solid development and risk management plans.

Evaluation of Socio-Economic Impact of City Bus Services in Developing Countries

Abstract City bus services encounter distinctive challenges in planning, maintaining and operating their services in developing countries. However, most of the studies are not structured in a simple manner and cannot find significant comparative information for socio-economic impact of city bus services. Therefore, this study presents a methodology for comparative evaluation of socio-economic impact of alternate city bus services. The methodology proposed in this study is also illustrated using a BRTS service in Bhopal city. It is expected that this study will be useful to take significant decisions before implementation of new services, and performance improvement of existing city bus services.

AN OBESE PATIENT “PORTRAIT” BY THE RESULTS OF OUTPATIENT CARDIOVASCULAR REGISTRY “PROFILE”

Aim. To assess the opportunities for registry methodology for comparative evaluation of clinical and anamnestic characteristics of cardiological patients with and without obesity; to draw a “portrait” of the obese patients, common for the cohort studied. Material and methods. The data analyzed from outpatient PROFILE registry during January 2011 to 31 august 2015: totally 1531 primary patient included, of those body mass index (BMI) collected in 1371. In 565 there was obesity of various grade, in 806 BMI was not higher than 30 kg/m2. The analysis done of retrospective data. Results. According to PROFILE results, obesity was found in 41,2% of patients, and those with obesity were younger (p=0,003). Among them men and women number was equa — 50,8% and 49,2%, resp. For the obese there were characteristic presence of dyspnea, oedema and low level of physical activity (p=0,0001). The obese had 1,5-3 times more common arterial hypertension, 2 type diabetes, chronic heart failure (p=0,0001), glucose tolerance disordered (p=0,001), dyslipidemia (p=0,012). Regardless that the obese took medications more commonly (p=0,006), they more rarely reached target level of blood pressure (BP), than those non obese (p<0,05). Almost nobody of those with obesity had prescribed therapy against overweight, except 73 patients that participated in 2 clinical trials. Conclusion. The data from outpatient registry PROFILE has confirmed the main information on obesity in cardiological patients: a “portrait” is drawn of the obese, cardiovascular risk factors (CRF) formulated, and comorbidities. The data demonstrates more difficult achievement of target BP in the obese with regular medication treatment, and highlights insufficient attention of physicians to diagnostics and especially treatment of obesity, as underestimation of the overweight problem by patients themselves.

New Methodology for Comparative Evaluation of the Abrasive Properties of Orthopedic Cables, Cut-Out Reciprocal Testing (CORT)

This work proposes a new testing methodology and an original testing bench to evaluate and compare the abrasive properties of orthopedic cables used in the surgical treatment of orthopedic disorders. This bench generates the cut-out reciprocal movement of a bone surrogate or biological bone specimen against a cable and it serves to evaluate the bone abrasion kinetics and the loss in cable tension when it cuts through bone. To validate the proposed methodology, a cut-out reciprocal testing study was performed using a conventional titanium orthopedic cable on both polyurethane foam bone surrogates and porcine vertebrae specimens. It was observed that the cutting rate and consequently the loss in cable tension were much lower for the porcine specimen than that observed for the bone surrogate given higher hardness of cortical bone compared to polyurethane. The methodology developed allows therefore comparative studies on the abrasiveness and on the ability to maintain the applied tension of orthopedic cables to be conducted.

A quality assessment index for evaluation of district e-governance websites

In this paper, a methodology for comparative evaluation of user perceived quality of websites is developed. This involves identification and scaling of factor categories relevant to users' perceived quality, proposing a website quality assessment index (WQAI) and using it for comparison of websites of selected districts of the state of Madhya Pradesh in India. Top websites were identified and evaluated by WQAI rankings and also through cluster analysis. It is found that: i) the best websites as identified by WQAI ranks match those obtained by cluster analysis, while there is a slight discrepancy in the lower order clusters; ii) websites of districts important because of population or industrial status may not be in the top user satisfaction category although they may have high ranks by other criteria. These findings will assist in the identification of strong and weak points of the websites and hence enable the governments to provide improved websites.

New Methodology for Comparative Evaluation of the Abrasive Properties of Orthopedic Cables, Cut-Out Reciprocal Testing (CORT)

New Methodology for Comparative Evaluation of the Abrasive Properties of Orthopedic Cables, Cut-Out Reciprocal Testing (CORT)

Comparative reliability analysis of the distribution areas of a water supply system and its application for an earthquake risk analysis

A methodology for comparative evaluation of the water supply reliability of distribution areas was presented. There are various factors which affect water supply reliability. As they cannot be quantified by the same standard, reliability of water distribution area was examined according to the value of each factor. These factors were organized into an index through the analytical hierarchy process. The reliability of each distribution area of the water supply system in the case study area was comparatively evaluated by the method. The problems and points necessary for improvement of the water supply reliability of each distribution area were indicated.

Asian Social Security Systems: A Comparative Analysis and Evaluation

This paper extends the comparative social security literature on Asia by incorporating an evaluative dimension. Whilst descriptive and analytical comparative social security research has a long tradition, only Kaim-Caudle has taken up the epistemological and methodological challenges of the comparative evaluation of social security systems, as distinct from programmes. The two-fold purpose of this paper is to describe the social security systems in 27 Asian countries and then to rank them using a comparative evaluation methodology that enables an assessment to be made of a country's statutory social security intention. The conclusion drawn is that the very best designed Asian social security systems are in post-socialist countries—particularly in Central Asia—all of which, however, face long-standing problems with respect to the conduct of their public administration or adequacy of their public finances. This means that the social security systems in these countries have at least become severely restricted because of war, natural disasters or severe economic dislocation. Outside Central Asia, the best-designed social security systems can be found in Japan (2nd), Hong Kong (7th), China (8th), the Philippines (9th) and Pakistan (10th). Outside Central Asia, only Japan and Malaysia have world-class designed social security programmes.

Sistemas de seguridade social na América Latina: uma avaliação ordinal

Este artigo amplia a literatura comparativa de sistemas de seguridade na América Latina através da incorporação da dimensão avaliativa. Seu objetivo é a classificação dos sistemas de seguridade nos 36 países latino-americanos utilizando uma metodologia que possibilita uma avaliação do modelo estatutário do sistema de cada país. A conclusão mostra que a região com os melhores modelos é a América Central, marginalmente à frente da América do Sul. Os melhores sistemas nacionais de seguridade na América Latina estão na Bolívia, Uruguai, Brasil, Nicarágua e Argentina. Observam-se programas de nível mundial no Brasil, Argentina, Chile, Peru e Nicarágua.

A Global Ranking of National Social Security Systems

There has been a very evident growth in descriptive and analytical comparative social security research over the past 40 years, but only Kaim‐Caudle has taken up the epistemological and methodological challenges of thecomparative evaluation of social security systems. The purpose of this paper is to rank social security systems in 172 countries using a comparative evaluation methodology that enables an assessment to be made of a country’s statutory social security intention. The conclusion drawn is that while seven of the ten best designed social security systems in the world can be found in western Europe, Australia has the best‐designed one.

African Social Security Systems: An Ordinal Evaluation

The purpose of this paper is to rank the social security systems in 45 African countries using a comparative evaluation methodology that enables an assess ment to be ma(le of a country's statutory social security intention. The conclusion drawn is that the spread of African social security system design standards are comparable to those of Latin American countries, although the poorest designed African systems are somewhat superior to their Latin American counterparts. The very best designed African social security systems are in North Africa: Tunisia (with its world-class family support program), Algeria and Libya, although Mauritius also stands out. Journal of Social Development in Africa Vol 15 No 1 2000, pp. 49-60