Article
Computer Science, Artificial Intelligence
Guanli Yue, Yanpeng Qu, Ansheng Deng, Qianyi Zhang
Summary: This paper proposes a method to reduce the impact of irrelevant features in nearest-neighbour classification by utilizing the learned knowledge of neural networks. Experimental results demonstrate that this method effectively reduces the influence of irrelevant features and enhances the performance of nearest-neighbour classification.
Article
Computer Science, Artificial Intelligence
Guo Feng Anders Yeo, Vural Aksakalli
Summary: A novel methodology based on simultaneous perturbation stochastic approximation (SPSA) for simultaneous feature selection and weighting for nearest neighbour (NN) learners is introduced in this study. Extensive computational experiments show that SPSA-FWS generally outperforms existing feature weighting algorithms and stands as a competitive new method for this task. Additionally, SPSA-FWS has attractive features allowing it to be used with any performance metric and any variant of nearest neighbour learners, and to be hybridised with other feature weighting methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Operations Research & Management Science
Pablo Aparicio-Ruiz, Elena Barbadilla-Martin, Jose Guadix, Pablo Cortes
Summary: This paper proposes a general framework to define indoor temperature based on a dynamic adaptive comfort algorithm. By learning the comfortable temperature with respect to running mean temperature, the method can determine the suitable indoor temperature range. The K-Nearest-Neighbour algorithm is shown to represent thermal comfort data patterns better than traditional solutions and is suitable for learning the thermal comfort area of a building and defining set-point temperature for HVAC systems.
ANNALS OF OPERATIONS RESEARCH
(2021)
Article
Instruments & Instrumentation
Min Song Seo, Ever Enrique Castillo Osorio, Hwan Hee Yoo
Summary: This study analyzes various fire risk factors in the urban area of Seoul and predicts their importance using machine learning techniques. The ignition condition is identified as the main factor in fire occurrence. The findings of this study can guide fire reduction and management measures in Seoul.
SENSORS AND MATERIALS
(2022)
Article
Computer Science, Information Systems
Dulakshi Santhusitha Kumari Karunasingha
Summary: The key statistical properties of RMSE and MAE estimators were derived for symmetric error distributions. An Approximate Root Normal Distribution (ARND) was developed to approximate the distribution of the RMSE estimator. The findings suggest that the choice between RMSE and MAE depends on the error distribution type, and the estimated RMSE/MSE ratio can identify the error distribution type.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Suvita Rani Sharma, Birmohan Singh, Manpreet Kaur
Summary: The binary versions of Rao algorithms are proposed for solving feature selection problems in Parkinson's disease datasets, optimizing the k parameter of the k-nearest neighbour classifier. The performance of these algorithms is evaluated through 30 independent runs with a 10-fold cross-validation procedure and compared with state of the art methods, with significance analysis conducted using the Friedman rank test.
Article
Chemistry, Multidisciplinary
Atul Kumar Mishra, Snehal Rajput, Meera Karamta, Indrajit Mukhopadhyay
Summary: Unlike conventional liquid electrolytes, solid-state electrolytes (SSEs) have gained attention in the domain of all-solid-state lithium-ion batteries (ASSBs) due to their safety features and better electrochemical stability. However, SSEs still face challenges such as poorer ionic conductivity and unstable physical characteristics. Machine learning has been used as a tool to predict new SSEs with improved properties for ASSBs.
Article
Multidisciplinary Sciences
Scott M. Robeson, Cort J. Willmott
Summary: When evaluating the performance of quantitative models, measures like mean squared error (MSE) and root mean squared error (RMSE) are commonly used to characterize dimensioned errors. However, absolute-value-based measures like mean absolute error (MAE) are more interpretable for quantifying average error. This study develops and demonstrates a decomposition of MAE into three submeasures, providing more straightforward information on the distribution of model error and proving to be preferable to comparable decompositions of MSE.
Article
Virology
Nasrullah Khan, Asma Arshad, Muhammad Azam, Ali Hussein Al-marshadi, Muhammad Aslam
Summary: The COVID-19 pandemic has had a significant impact globally, leading researchers and governments to seek ways to control the spread of the virus and accelerate the treatment process. To effectively predict the development trend of the pandemic and reduce losses, they have developed a forecasting model to help researchers, governments, and others devise response strategies.
JOURNAL OF MEDICAL VIROLOGY
(2022)
Article
Green & Sustainable Science & Technology
J. N. Chandra Sekhar, Bullarao Domathoti, Ernesto D. R. Santibanez Gonzalez
Summary: Electrified transportation systems are rapidly emerging worldwide, contributing to the reduction of carbon emissions and global warming. Battery remaining useful life (RUL) prediction is crucial for cost reduction and improving system reliability and efficiency. Existing prediction approaches for battery performance evaluation are unsatisfactory. This study aims to enhance prediction accuracy and robustness using selected machine learning algorithms.
Article
Computer Science, Information Systems
Mohamad Abou Houran, Mohamed H. Essai Ali, Adel B. Abdel-Raman, Eman A. Badry, Alaaeldien Hassan, Hany A. Atallah
Summary: This paper suggests improving the performance of Deep Learning Long Short-Term Memory (DLLSTM) structures by using robust loss functions and creating new classification layers. The effectiveness of the suggested DLLSTM classifier was examined using three loss functions (Crossentropy, MAE, SSE) for two different applications. The results show that the suggested classifier with SSE loss function outperforms others and the suggested activation functions are more accurate than the tanh function.
Article
Construction & Building Technology
Kun Gao, Aoyong Li, Yang Liu, Jorge Gil, Yiming Bie
Summary: Understanding the mode substitution of dockless bike sharing (DLBS) in relation to other transport modes is crucial for assessing their impact and planning for improvements. This study utilizes multi-modal route planning techniques, transaction data of bike-sharing, and travel behavior modeling to analyze the mode substitution of DLBS at the trip level. The study also employs interpretable machine learning to uncover the effects of built environment factors on mode substitution patterns. The results show heterogeneity in the probabilities of DLBS replacing different transport modes, and built environment factors play a role in explaining these variations.
SUSTAINABLE CITIES AND SOCIETY
(2023)
Article
Green & Sustainable Science & Technology
Adnan Yousaf, Rao Muhammad Asif, Mustafa Shakir, Ateeq Ur Rehman, Fawaz Alassery, Habib Hamam, Omar Cheikhrouhou
Summary: This paper presents a novel and improved technique to forecast electricity prices by considering various data and using different algorithms to reduce forecasting error. The proposed integration strategy computes the mean absolute percentage error (MAPE), achieving an overall improvement percentage of 9.24% valuable in price forecasting for the energy management system (EMS).
Article
Economics
Kun Gao, Ying Yang, Aoyong Li, Junhong Li, Bo Yu
Summary: This study proposes an innovative method for quantifying the economic benefits of free-floating bike-sharing systems (FFBS) and conducts an empirical analysis in Shanghai to estimate the saved travel time, cost, and economic benefit due to using FFBS per trip. The results show that using FFBS can save 9.95 minutes, 3.64 CNY, and provide an economic benefit of 8.68 CNY-eq per trip, with annual savings of 17.665 billion min, 6.463 billion CNY, and 15.410 billion CNY-eq in Shanghai. Additionally, the study quantitatively examines the relationship between economic benefits from FFBS and built environment factors in different urban contexts using Multiple Linear Regression.
TRANSPORTATION RESEARCH PART A-POLICY AND PRACTICE
(2021)
Article
Computer Science, Information Systems
Hongbo Zhang, Xin Gao, Jixiang Du, Qing Lei, Lijie Yang
Summary: The study proposes a photomosaic image generation method that minimizes both global and local errors through tile selection and weight coefficient adjustment, resulting in images with a more artistic effect.
FRONTIERS OF COMPUTER SCIENCE
(2021)
Article
Computer Science, Information Systems
Malliga Subramanian, Jaehyuk Cho, Veerappampalayam Easwaramoorthy Sathishkumar, Obuli Sai Naren
Summary: Cancer is the second leading cause of death globally, with one in six deaths attributed to it. Early detection improves the chances of survival, and the use of Artificial Intelligence (AI) for automated cancer detection can help evaluate more cases in less time.
Article
Management
Li Yang, V. E. Sathishkumar, Adhiyaman Manickam
Summary: This paper proposes an integrated deep reinforcement learning-based logistics management model (DELLMM) to increase and optimize logistic distribution. An optimization approach can be used in inventors and price control applications. The research methodology gives the fundamentals of information retrieval and the scope of blockchain integration. The experimental results show that DELLMM improves logistics management and optimized distribution compared to other methods with the highest operability of 94.35%, latency reduction of 97.12%, efficiency of 98.01%, trust enhancement of 96.37%, and sustainability of 97.80%.
INTERNATIONAL JOURNAL OF INFORMATION SYSTEMS AND SUPPLY CHAIN MANAGEMENT
(2023)
Article
Computer Science, Artificial Intelligence
Neelakandan Subramani, Sathishkumar Veerappampalayam Easwaramoorthy, Prakash Mohan, Malliga Subramanian, Velmurugan Sambath
Summary: Twitter, Instagram and Facebook are rapidly expanding and reporting daily news, social activities and actual occurrences. Social network analysis (SNA) research faces ethical challenges due to technology advances and increasing ethics regulation. This study investigates how influencer content generates interactions and develops a framework for identifying users with the ability to influence others.
BIG DATA AND COGNITIVE COMPUTING
(2023)
Article
Multidisciplinary Sciences
Malliga Subramanian, Veerappampalayam Easwaramoorthy Sathishkumar, Jaehyuk Cho, Kogilavani Shanmugavadivel
Summary: COVID-19, a global pandemic, has a high mortality rate. CT scans can assist in diagnosing and monitoring the disease, but visual inspection is time-consuming. This study utilizes a Convolution Neural Network (CNN) with transfer learning and integrates Learning without Forgetting (LwF) to enhance the model's generalization capabilities. The wide ResNet model with LwF method achieves superior performance in classifying original and delta-variant datasets.
SCIENTIFIC REPORTS
(2023)
Article
Ecology
Veerappampalayam Easwaramoorthy Sathishkumar, Jaehyuk Cho, Malliga Subramanian, Obuli Sai Naren
Summary: This study investigates fire/smoke detection from images using AI-based computer vision techniques. Transfer learning is implemented on pre-trained models to reduce training time and complexity. The Xception model performs well with LwF and achieves high accuracy on both the new and original datasets.
Article
Engineering, Multidisciplinary
V. E. Sathishkumar, A. G. Ramu, Jaehyuk Cho
Summary: This research study utilized machine learning techniques to effectively remove hazardous substances like azo dyes and nitrophenols from drinking water using the catalyst PdO-NiO. The results showed that the XGB algorithm performed best with 4-NP and DNP, the RF algorithm performed best with TNP, MB, and RHB, and the SVM algorithm performed best with MO.
ALEXANDRIA ENGINEERING JOURNAL
(2023)
Article
Engineering, Multidisciplinary
S. Anbukkarasi, D. Elangovan, Jayalakshmi Periyasamy, V. E. Sathishkumar, S. Sree Dharinya, M. Sandeep Kumar, J. Prabhu
Summary: Transliteration is the process of mapping the characters of one language to those of another language based on phonetics. This process is particularly important in India, where people speak a variety of languages and may struggle to read different scripts. Transliteration plays a crucial role in various Natural Language Processing applications, including information retrieval, machine translation, and speech recognition. While transliteration works have been carried out in languages like Japanese, Chinese, and English, there is limited research on Indian languages, especially Tamil. This paper focuses on the transliteration of Unicode Tamil characters using a phonetics-based forward list processing method, which shows promising results.
INTERNATIONAL JOURNAL OF RELIABILITY QUALITY AND SAFETY ENGINEERING
(2023)
Review
Cell Biology
Kogilavani Shanmugavadivel, V. E. Sathishkumar, Jaehyuk Cho, Malliga Subramanian
Summary: This article reviews methods and techniques for early detection of Alzheimer's Disease and provides a comprehensive analysis of AD diagnosis datasets. The research findings are important for improving the accuracy of Alzheimer's Disease detection.
AGEING RESEARCH REVIEWS
(2023)
Article
Green & Sustainable Science & Technology
Malliga Subramanian, Jaehyuk Cho, Sathishkumar Veerappampalayam Easwaramoorthy, Akash Murugesan, Ramya Chinnasamy
Summary: This study compares the performance of various time series and machine learning algorithms for predicting bike demand and finds that GRU algorithm performs the best. ARIMA and SARIMA models produce less accurate predictions, likely due to their assumptions of linearity and stationarity in the data.
Article
Multidisciplinary Sciences
Shaik Jakeer, Sathishkumar Veerappampalayam Easwaramoorthy, Seethi Reddy Reddisekhar Reddy, Hayath Thameem Basha
Summary: This study presents a novel implementation of an intelligent numerical computing solver using an MLP feed-forward backpropagation ANN and the Levenberg-Marquard algorithm to interpret the Cattaneo-Christov heat flux model. The effect of entropy production and melting heat transfer on the ferrohydrodynamic flow of the Fe3O4-Au/blood Powell-Eyring hybrid nanofluid is demonstrated. The artificial neural network model is used for data selection, network construction, training, and evaluation, with various physical factors impacting variables such as velocity, temperature, entropy generation, friction coefficient, and heat transfer rate.
Article
Engineering, Multidisciplinary
V. E. Sathishkumar, R. Vadivel, Jaehyuk Cho, Nallappan Gunasekaran
Summary: In this paper, the finite-time dissipativity analysis of Markovian jump-delayed neural networks (MJDNNs) is studied. Less conservative results for extended dissipativity conditions are established for delayed MJDNNs. An appropriate Lyapunov-Krasovskii functional (LKF) with novel inequality, the composite slack-matrix-based integral inequality (CSMBII), is used to achieve this. Sufficient conditions including CSMBII are employed to derive a delay-dependent finite-time dissipativity condition in terms of linear matrix inequalities (LMIs), which are used to formulate the finite dissipativity condition for the delayed MJNNs. Numerical examples confirm the utility of the suggested approach, including a real-world application of the benchmark problem associated with the designed MJDNNs.
ALEXANDRIA ENGINEERING JOURNAL
(2023)
Article
Mathematics
Shaik Jakeer, Seethi Reddy Reddisekhar Reddy, Sathishkumar Veerappampalayam Easwaramoorthy, Hayath Thameem Basha, Jaehyuk Cho
Summary: This study investigates the importance of induced magnetic fields and double-diffusive convection in the radiative flow of Carreau nanofluid through three different geometries. The fluid transport equations were simplified using self-similarity variables and solved using the Runge-Kutta-Fehlberg method. The study demonstrates how various dynamic factors influence the fluid's transport characteristics through graphical representations.
Article
Computer Science, Information Systems
S. Anbukkarasi, Veerappampalayam Easwaramoorthy Sathishkumar, C. R. Dhivyaa, Jaehyuk Cho
Summary: Recognizing text from nature scene images and videos is a challenging task due to their complex features and variations. However, text recognition is highly useful in various applications. This research paper proposes a model that combines Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) to successfully detect and recognize text characters in natural images. The experimental results demonstrate that the proposed model outperforms other methods on multiple datasets.
Article
Mathematics, Applied
S. Neelakandan, Sathishkumar Veerappampalayam Easwaramoorthy, A. Chinnasamy, Jaehyuk Cho
Summary: It has been shown that fuzzy systems are useful for classification and regression, but they are mostly used in controlled environments. An image clustering technique using color, texture, and shape information is developed for content-based picture retrieval in large image datasets. The challenge of labeling a large number of photos is addressed by using unsupervised learning, specifically the K-means clustering algorithm. In comparison to fuzzy c-means clustering, K-means clustering has better performance in lower-dimensional space resilience and initialization resistance. The dominant triple HSV color space is a perceptual color space composed of saturation (S), hue (H), and value (V), which are closely related to human color perception. A deep learning technique called RBNN is built using Gaussian function, fuzzy adaptive learning control network (FALCN), clustering, and radial basis neural network to achieve image segmentation and feature extraction. The suggested FALCN fuzzy system is excellent at clustering images and extracting image properties. Traditional fuzzy network systems tend to have redundant output neurons when receiving noisy input. Finally, random convolutional weights are used to extract features from unlabeled data.
Article
Engineering, Multidisciplinary
V. E. Sathishkumar, Yongyun Cho
Summary: In classical regression theory, fitting a single function model to a data set is a complex and unreliable process in complex and noisy domains. To overcome these difficulties, piecewise regression models and proper feature selection are proposed in this paper. The hybridization of Elephant Herding Optimization (EHO) and minimum Redundancy and Maximum Relevance (mRMR) is used for feature selection to improve regression problems. The results demonstrate the effectiveness of CUBIST and mRMR-EHO feature selection in various datasets and indicate that it can be used as an effective tool for predictive data modeling.
TEHNICKI VJESNIK-TECHNICAL GAZETTE
(2023)
