Selected Machine Learning Algorithms Research Articles

An AIoT-based hydroponic system for crop recommendation and nutrient parameter monitorization

Advancements in technology have revolutionized various sectors, including agriculture, which serves as the backbone of many economies, particularly in Asian countries. The integration of new technologies and research has consistently aimed to enhance cultivation rates and reduce reliance on manual labor. Two key technologies, Artificial Intelligence (AI) and the Internet of Things (IoT), have emerged as pivotal tools in automating processes, providing recommendations, and monitoring agricultural activities to optimize results. While traditional soil cultivation has been the preferred method, the increasing urbanization trend necessitates alternative approaches such as hydroponics, which replaces soil with water as the medium for crop cultivation. Having many significant advantages, hydroponics serves a crucial role in achieving efficient space utilization. To get a higher density of plants in a confined area hydroponic approach provides water, nutrients and other essential elements directly to the plant's root. To utilize the hydroponic system more effectively, our proposed method, integrating AI and IoT helps to provide suitable crop recommendations, monitor the parameters of the plants and also suggest the necessary changes required for gaining optimal parameters. To ensure optimal resource allocation and maximize yields we have used machine learning models and trained them to recommend suitable crops from the given parameters and also refer to the changes in parameters that are needed for better plant growth. We have used the crop recommendation dataset from the Indian Chamber of Food and Agriculture to train our proposed machine-learning model. Our selected machine learning algorithms to predict the best crops are Random forests, Decision trees, SVM, KNN, and XGBoost. Our research combines AI and IoT with hydroponic systems to streamline crop recommendations, automate monitoring processes, and provide real-time guidance for optimized cultivation. Among them, the Random forest algorithm outperformed other algorithms with an accuracy of 97.5%.

Air Quality Prediction usingKNN and LSTM

The project titled "Air Quality Prediction Using KNN and LSTM" endeavors toaddress the critical issue of air pollutionthrough the application of advanced computational techniques. The project aimsto develop a robust predictive model that can forecast air quality levels based on historical data, meteorological parameters, and relevant environmental features. Leveraging machine learning algorithms such as regression, decision trees, or neural networks, the project seeks to analyze complex relationships within the data and enhance the accuracy of air qualitypredictions. The methodology involves the collection and preprocessing of extensive datasets encompassing pollutant concentrations, weather conditions, and geographicalinformation. The selected machine learning algorithms will be trained on this data to recognize patterns and correlations, enabling the model to make accurate predictions. The project also explores the integration of real- time data streams, satellite imagery, and sensor networks to improve the responsiveness of the predictive model.

Prototyping a Reusable Sentiment Analysis Tool for Machine Learning and Visualization

Evaluating customer satisfaction is very significant in all organizations to get the perspective of users/customers/stakeholders on products and/or services. Part of the data obtained during the evaluation are observations and comments of respondents and these are very rich in insights as they provide information on the strengths as well as the areas needing improvement. As the volume of textual data increases, the difficulty of analyzing them manually also increases. With these concerns, text analytics tools should be used to save time and effort in analyzing and interpreting the data. The textual data being processed in sentiment analysis problems vary in so many ways. For instance, the context of textual data and the language used vary when data are sourced from different locations and areas or fields. Thus, machine learning was utilized in this study to customize text analysis depending on the context and language used in the dataset. This research aimed to produce a prototype that can be used to explore three vectorization techniques and selected machine learning algorithms. The prototype was evaluated in the context of features for the application of machine learning in sentiment analysis. Results of the prototype development and the feedback and suggestions during the evaluation were presented. In future work, the prototype shall be improved, and the evaluators' feedback will be considered.

A comparative study of selected machine learning algorithms for electrical impedance tomography

A comparative study of selected machine learning algorithms for electrical impedance tomography

Étude comparative d'algorithmes d'apprentissage artificiel pour la reconnaissance faciale

Background: The fundamental need for authentication and identification of humans using their physiological, behavioral or biological characteristics, continues to be applied extensively to secure localities, property, financial transactions, etc. Biometric systems based on face characteristics, continue to attract the attention of researchers, major public and private services. In the literature, many methods have been deployed by different authors. The best performance must be found in order to be able to recommend the most effective method. So, the main objective of thisarticle is to make a comparative study of different existing techniques.Methods: A biometric system is generally composed of four stages: acquisition of facial images, preprocessing, extraction of characteristics and finally classification. In this work, the focus is on machine learning algorithms for classification. These algorithms are: Support Vector Machines (SVM), Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Random Forests (RF), Logistic Regression (LR), Naive Bayesian Classification (NB: Naive Bayes’ Classifiers) and deep learning techniques such as Convolutional Neural Networks (CNN). The comparison criterion is the average performance, calculated using three performance measures: recognition rate, confusion matrix, and the Area Under Receiver Operating Characteristic (ROC) curve.Results: Based on this criterion, the performance comparison of selected machine learning algorithms, shows that CNN is the best, with an average performance of 100.00% On ORL face database. However, on the YALE database, classical algorithms such as artificial neural networks have obtained the best performances, the highest being a rate of 100%.Discussion: Deep learning techniques are very efficient in image classification as proven by the results on the ORL database. However, their inefficiency on YALE face database is due to the small size of this database which is inappropriate for some deep learning algorithms. But this weakness can be corrected by image augmentation techniques. The comparison of these results with existing state-of-the-art methods is nearly the same. Authors achieved performances of 94.82%, 95.79%, 96.15%, 96.44%, 97.27%, 98.52% and 98.95% for NB, KNN, RF, LR, ANN, SVM and CNN classifiers, respectively. Finally, in depth discussion, it is concluded that between all these approaches which are useful in face recognition, the CNN is the best classification algorithm.

Machine learning facilitates connections between soil thermal conductivity, soil water content, and soil matric potential

Properly understanding the strongly coupled thermal and hydraulic properties of porous media is crucial to interests in different fields, e.g., agriculture, building physics and meteorology. Previous studies generally consider the thermal and hydraulic properties independently, specifically, the soil thermal conductivity curve and the soil water retention characteristics curve as methods to connect such process are unavailable or untested. In an effort to test methods based on machine learning (ML) to effectively integrate the soil thermal conductivity (λ), matric potential (ψ) and water content (θ), 2328 measurements from 43 soils were compiled using ML algorithms. Five ML algorithms, including random forest-RF, gradient boosting decision tree-GBDT, support vector machine-SVM, multiple linear regression-LR and backpropagation neural network-BPNN, based on the balanced dataset with synthetic minority over-sampling technique were trained, validated and tested to establish the λ(ψ) and λ(θ) relationships. Results demonstrate that, ML algorithms perform better than empirically based methods to optimize the thermal and hydraulic relationship. The selection of ML algorithms has a significant influence on the coupling relationship and should be taken into account. GBDT performed superior for small soil hydrothermal datasets (Testing set: Nash-Sutcliff efficiency-NSE = 0.9999 and Root Mean Square Error-RMSE = 0.0144 W m−1 K−1; independent verification: NSE = 0.7122 and RMSE = 0.2726 W m−1 K−1), while BPNN is limited by the amount of data with NSE = 0.9970, RMSE = 0.0790 W m−1 K−1 of Testing set.

Identifying and Predicting Muslim Community Funeral Funding Protocols

This research aims to understand funeral poverty among the Muslim community in Malaysia by using Machine Learning algorithms. Generally, funeral poverty in Malaysia can be described as funerals in Malaysia being exorbitantly priced, inflicting a disproportionate amount of financial hardship on the poor. As the death rate continues to climb year after year, we may conclude that funeral fees are an unavoidable risk. This research problem was inspired by Sustainable Development Goal number 1 and 3 where the goal of this initiative is to increase awareness about the need of having funeral expenditures and how to prevent funeral poverty. Previous works have shown promising results. However, they are not conclusive enough as they lack predicting capability that Machine Learning can offer. Selected Machine Learning algorithms such as Decision Tree, Random Forest, and Naïve Bayes were used to classify the people that will go through funeral poverty based on a selected dataset and a survey conducted. The research methodology is as follows: 1. Collecting the data, 2. Pre-Process the data, 3. Exploratory data analysis, 4. Selecting the feature, 5. Modeling, 6. Evaluation. The results showed that the accuracy value for the selected algorithm is 0.95, 0.95, and 0.95 for the awareness and 0.8, 0.8, and 0.75 for B40 Status in survey datasets. With a larger taxonomy and more extensive, diverse sets of data, these figures are expected to improve. Linear Regression has been applied to help understand the features of a death rate. Both outcomes serve distinct reasons. Because the individuals who would experience funeral poverty are in lower-income neighborhoods, it was necessary to classify impoverished communities in order to anticipate funeral poverty. Following that, the findings of what causes death to rise aid in understanding how the B40 community has been impacted by funeral poverty. In the future, this study will aid future research on funeral poverty on a larger scale. When the two datasets are integrated rather than separated, better results can be obtained. The accuracy of the findings may also be improved by employing a wider range of Machine Learning methods and deep learning implementation. A more coherent and detailed model for forecasting funeral poverty can be developed.

Prediction of robo-advisory acceptance in banking services using tree-based algorithms.

The banking sector is increasingly recognising the need to implement robo-advisory. The introduction of this service may lead to increased efficiency of banks, improved quality of customer service, and a strengthened image of banks as innovative institutions. Robo-advisory uses data relating to customers, their behaviors and preferences obtained by banks from various communication channels. In the research carried out in the work, an attempt was made to obtain an answer to the question whether the data collected by banks can also be used to determine the degree of consumer interest in this type of service. This is important because the identification of customers interested in the service will allow banks to direct a properly prepared message to a selected group of addressees, increasing the effectiveness of their promotional activities. The aim of the article is to construct and examine the effectiveness of predictive models of consumer acceptance of robo-advisory services provided by banks. Based on the authors' survey on the use of artificial intelligence technology in the banking sector in Poland, in this article we construct tree-based models to predict customers' attitudes towards using robo-advisory in banking services using, as predictors, their socio-demographic characteristics, behaviours and attitudes towards modern digital technologies, experience in using banking services, as well as trust towards banks. In our study, we use selected machine learning algorithms, including a decision tree and several tree-based ensemble models. We showed that constructed models allow to effectively predict consumer acceptance of robo-advisory services.

Performance Assessment of Machine Learning Techniques for Gully Erosion Mapping in South-East Nigeria

Soil erosion is a serious environmental hazard affecting southeast Nigeria. The rate of erosion has continued to increase with devastating impact and has led to land degradation (gullies). To assess the socio-economic and environmental implications of gully erosion and to develop management plans to deal with them, quantitative and qualitative data on soil erosion rates at regional scales are needed. Hence, this study aimed at investigating the spatial and temporal variations of gully erosion impacted areas of South East Nigeria using machine learning techniques to develop a suitable approach for accurately modelling gully erosion impact in South East, Nigeria. Its objectives are to: acquire, process, and perform object-based gully mapping analysis with Kompsat-3 imagery of South East Nigeria, Nigeria using three machine learning algorithms (Linear discriminant analysis, Support Vector Machine, and Artificial Neural Network); compare the performance of the three selected machine learning algorithms using statistical techniques and determine the best suited for gully mapping within the study area; determine the pattern and trend of Gully Erosion in South East, Nigeria between 2010 and 2020, using the best-suited machine learning algorithm; and predict the future gully development pattern for the next 10 years. The methodology involved the acquisition and use of KOMPSAT-3 imagery, identification of gully erosion impacted areas, and topographic maps. Image preprocessing, development of classification scheme, object-based image analysis, trend analysis, and pattern of gully development and expansion. Accuracy assessment and gully development prediction to 2030 as data analysis. The results showed consistent identification of eight class features, including waterbody, dense vegetation, agricultural land, bare ground, built-up area, sand, savannah, and gully erosion. The performance evaluation of the three techniques revealed promising outcomes. LDA achieved an accuracy of 0.759, while SVM demonstrated a higher accuracy of 0.853 and balanced sensitivity and specificity. ANN showed good performance with an accuracy of 0.790. Based on its superior accuracy and balanced performance, SVM was selected for mapping the spatial trend of gully erosion in the study area.

Automated Machine Learning: A Comprehensive Survey and Framework for Advancements

Abstract: Machine learning has become an indispensable tool in numerous domains, and the selection of an appropriate machine learning algorithm can significantly impact the success of a project. To address the challenge of algorithm selection, we present an Automated Machine Learning (AutoML) platform designed to empower users with the ability to effortlessly compare the performance of two selected machine learning algorithms on their datasets. This platform offers a user-friendly interface, guiding users through the entire process, from uploading their datasets to selecting algorithms and making informed decisions. The platform provides a streamlined workflow that allows users to easily upload datasets of their choice, select two machine learning algorithms from a pre-defined set, and compare their performance based on a variety of evaluation metrics. Users can visualize and analyze the results to gain insights into how different algorithms impact their data. Additionally, the platform offers assistance in identifying the most suitable algorithm for achieving optimal results. With the rapid evolution of machine learning techniques, this AutoML platform addresses the need for a user-friendly, efficient, and accessible tool to assist practitioners and researchers in algorithm selection. By simplifying the comparative analysis of machine learning models, this platform empowers users to make informed decisions that enhance the effectiveness of their data-driven projects.

Machine learning algorithm and neural network architecture for optimization of pharmaceutical and drug manufacturing industrial effluent treatment using activated carbon derived from breadfruit (Treculia africana)

In a recent development, attention has shifted to the application of artificial intelligence for the optimization of wastewater treatment processes. This research compared the performances of the machine learning (ML) model: random forest, decision tree, support vector machine, artificial neural network, convolutional neural network, long-short term memory, and multiple linear regressors for optimization in effluent treatment. The training, testing, and validation datasets were obtained via the design of an experiment conducted on the removal of total dissolved solids (TDS) from pharmaceutical effluent. The breadfruit-activated carbon (BFAC) adsorbent was characterized using scanning electron microscopy and X-ray diffraction techniques. The predictive capacity of an ML algorithm, and neural network architecture implemented to optimize the treatment process using statistical metrics. The results showed that MSE ≤ 1.68, MAE ≤ 0.95, and predicted-R2 ≥ 0.9035 were recorded across all ML. The ML output with minimum error functions that satisfied the criterion for clean discharge was adopted. The predicted optimum conditions correspond to BFAC dosage, contact time, particle size, and pH of 2.5 mg/L, 10 min, 0.60 mm, and 6, respectively. The optimum transcends to a reduction in TDS concentration from 450 mg/L to a residual ≤ 40 mg/L and corresponds to 90% removal efficiency, indicating ± 1.01 standard deviation from the actual observation practicable. The findings established the ML model outperformed the neural network architecture and affirmed validation for the optimization of the adsorption treatment in the pharmaceutical effluent domain. Results demonstrated the reliability of the selected ML algorithm and the feasibility of BFAC for use in broad-scale effluent treatment.

Coronary vascular function in females with history of hypertensive disorders of pregnancy an oxygenation-sensitive cardiovascular magnetic resonance imaging study

Abstract Background Hypertensive disorders of pregnancy (HDP), including gestational hypertension and preeclampsia, have been associated with increased risk for cardiovascular disease decades after HDP pregnancy. HDP is thought to unmask pre-existing cardiovascular risk in the mother; and in addition HDP triggers a cascade of inflammatory and oxidative stress which leads to vascular endothelial dysfunction and abnormalities in cardiac structure and function evident early post-partum that may persist sub-clinically. Oxygenation-sensitive cardiovascular magnetic resonance (OS-CMR) is an emerging approach that can identify dynamic changes of myocardial oxygenation as a marker for coronary vascular function. The objective of this study was to assess coronary vascular function in females with history of HDP using OS-CMR. Methods We analyzed data from two short axis OS-CMR slices in 20 females with history of HDP, including gestational hypertension, preeclampsia and HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets) and 20 healthy controls with history of normal pregnancy acquired on a 1.5T MRI scanner. Images were analyzed using a prototype tissue oxygenation module. The images were normalized for signal intensity to both ventricular blood pools. We used both traditional statistics and feature selection machine learning algorithms to select for predictive biomarkers in the dataset. Finally, we used a random forest classifier model for disease classification. We report area under the receiver operating curve (AUROC) as a metric for diagnostic accuracy. Results Global breathing-induced myocardial oxygenation reserve (B-MORE) of the mid left ventricular slice was significantly reduced in HDP females (-1.6 ± 14.8) when compared to healthy controls (11.3 ± 17.2, p=0.03), when normalized to ventricular blood pools (Figure 1A). Our feature selection algorithm selected the signal intensity during end-systole normalized to the series as the most predictive biomarker. Our model, which integrated data from >3,300 discrete data points per participant, correctly differentiated healthy controls from women with HDP history with a 90% ± 3% AUROC accuracy, indicating high sensitivity and specificity (Figure 1B). Conclusion Compared to healthy controls, females with history of HDP have abnormal coronary vascular function, as shown by reduced myocardial oxygenation on OS-CMR. Combining advanced machine learning approaches can further enhance the potential application of OS-CMR to risk stratify females with adverse pregnancy outcomes susceptible to future cardiovascular disease.

An extension of iStar for Machine Learning requirements by following the PRISE methodology

The rise of Artificial Intelligence (AI) and Deep Learning has led to Machine Learning (ML) becoming a common practice in academia and enterprise. However, a successful ML project requires deep domain knowledge as well as expertise in a plethora of algorithms and data processing techniques. This leads to a stronger dependency and need for communication between developers and stakeholders where numerous requirements come into play. More specifically, in addition to functional requirements such as the output of the model (e.g. classification, clustering or regression), ML projects need to pay special attention to a number of non-functional and quality aspects particular to ML. These include explainability, noise robustness or equity among others. Failure to identify and consider these aspects will lead to inadequate algorithm selection and the failure of the project. In this sense, capturing ML requirements becomes critical. Unfortunately, there is currently an absence of ML requirements modeling approaches. Therefore, in this paper we present the first i* extension for capturing ML requirements and apply it to two real-world projects. Our study covers two main objectives for ML requirements: (i) allows domain experts to specify objectives and quality aspects to be met by the ML solution, and (ii) facilitates the selection and justification of the most adequate ML approaches. Our case studies show that our work enables better ML algorithm selection, preprocessing implementation tailored to each algorithm, and aids in identifying missing data. In addition, they also demonstrate the flexibility of our study to adapt to different domains.

Comparison of Selected Machine Learning Algorithms in the Analysis of Mental Health Indicators

Machine learning is increasingly being used to solve clinical problems in diagnosis, therapy and care. Aim: the main aim of the study was to investigate how the selected machine learning algorithms deal with the problem of determining a virtual mental health index. Material and Methods: a number of machine learning models based on Stochastic Dual Coordinate Ascent, limited-memory Broyden–Fletcher–Goldfarb–Shanno, Online Gradient Descent, etc., were built based on a clinical dataset and compared based on criteria in the form of learning time, running time during use and regression accuracy. Results: the algorithm with the highest accuracy was Stochastic Dual Coordinate Ascent, but although its performance was high, it had significantly longer training and prediction times. The fastest algorithm looking at learning and prediction time, but slightly less accurate, was the limited-memory Broyden–Fletcher–Goldfarb–Shanno. The same data set was also analyzed automatically using ML.NET. Findings from the study can be used to build larger systems that automate early mental health diagnosis and help differentiate the use of individual algorithms depending on the purpose of the system.

Aliro: an automated machine learning tool leveraging large language models.

Biomedical and healthcare domains generate vast amounts of complex data that can be challenging to analyze using machine learning tools, especially for researchers without computer science training. Aliro is an open-source software package designed to automate machine learning analysis through a clean web interface. By infusing the power of large language models, the user can interact with their data by seamlessly retrieving and executing code pulled from the large language model, accelerating automated discovery of new insights from data. Aliro includes a pre-trained machine learning recommendation system that can assist the user to automate the selection of machine learning algorithms and its hyperparameters and provides visualization of the evaluated model and data. Aliro is deployed by running its custom Docker containers. Aliro is available as open-source from GitHub at: https://github.com/EpistasisLab/Aliro.

Prediction of the Type of Organizational Culture Using Machine Learning Approach

The main purpose of the article is to develop a method for predicting the type of organizational culture using machine learning. The method is to be based on the popular OCAI survey by Cameron and Quinn (2022). To achieve the main goal of the study, selected machine learning algorithms were utilized. The developed method primarily reduces cognitive effort both on the part of respondents to traditional organizational culture survey sheets and on the part of researchers who analyze the results of traditional organizational culture diagnosis sheets. The reduction in cognitive effort on the respondents' side is due to the fact that they do not have to distribute 100 points for each of the six questions for each type of organizational culture. In the developed method, the only task of the respondents is to indicate the dominant type of organizational culture in each of the six criteria (questions in the survey). An additional purpose of the article is to present machine learning as a useful quantitative instrument for solving problems that occur in the theory and practice of management and quality sciences.

The use of machine-learning methods for post-earthquake building usability assessment: A predictive model for seismic-risk impact analyses

The assessment of building usability in the aftermath of an earthquake is mostly aimed at post-event emergency management, but it is also valuable for the planning of risk-reduction policies. In the seismic risk assessment field, the development of suitable consequence functions that correlate physical damage to usability and serviceability of structures is crucial to evaluate the expected social and economic losses in a region of interest. Predictive models for usability classification generally are calibrated on empirical data and provide the probability of loss of usability as function of the intensity measure, the building type and the severity of damage attained by the structure. Exploiting the large amount of data available in Italy, a decision tree-based approach is proposed in this study to assess post-earthquake usability of ordinary buildings. Thanks to its high interpretability coupled with reasonable predictive capability _, the selected machine learning algorithm allows investigation of the structural parameters that have a significant impact on building usability, while also accounting for the traditionally neglected uncertainty of subjective decisions. Finally, to show the potential of the proposed usability consequence models, a large-scale risk analysis is carried out to evaluate the spatial distribution of expected building-usability losses over time.

Effectiveness of selected AI models in predicting victims of terrorist attacks

Terrorism continues to be a problem for many countries around the world. The article compares the effectiveness of selected machine learning algorithms in predicting the victims of terrorist attacks in order to answer the question of whether they can serve as one of the anti-terrorist tools. An exploratory data analysis was carried out, and selected trends and characteristics of terrorist attacks were discussed. Some measures for evaluating the classification algorithms used in the study are presented, and potential directions for further research are indicated.

Stock picking with machine learning

AbstractWe analyze machine learning algorithms for stock selection. Our study builds on weekly data for the historical constituents of the S&P500 over the period from January 1999 to March 2021 and builds on typical equity factors, additional firm fundamentals, and technical indicators. A variety of machine learning models are trained on the binary classification task to predict whether a specific stock outperforms or underperforms the cross‐sectional median return over the subsequent week. We analyze weekly trading strategies that invest in stocks with the highest predicted outperformance probability. Our empirical results show substantial and significant outperformance of machine learning‐based stock selection models compared to an equally weighted benchmark. Interestingly, we find more simplistic regularized logistic regression models to perform similarly well compared to more complex machine learning models. The results are robust when applied to the STOXX Europe 600 as alternative asset universe.

NOVEL METHODOLOGY FOR REAL-TIME STRUCTURAL ANALYSIS ASSISTANCE IN CUSTOM PRODUCT DESIGN

Mass-customization is related to optimizing the balance between flexibility, strongly required by the customer-focused industries and manufacturing efficiency, which is critical for market competitiveness. In the conventional industries, the process of designing, validating and manufacturing a product is long and expensive. Some of the common approaches for addressing those issues are parametric product modeling and Finite Element Analysis (FEA). However, the costs involved are still relatively high because of the very special expertise needed and the cost of the specialized software. Also, the specific design of the product cannot be validated in a real-time, which often leads to making hard compromises between the specific customer requirements and the structural properties of the product in its exploitation. In this paper, we propose the novel methodology for real-time structural analysis assistance for custom product design. We introduce the concept of so-called compiled FEA model, a Machine Learning (ML) model, consisting of dataset of characteristic product parameters and associated physical quantities and properties, selected ML algorithms and the sets of associated hyperparameters. A case study of creating a compiled FEA model for the case of internal orthopedic fixator is provided.