Use Of Machine Learning Algorithms Research Articles

Customer data prediction and analysis in e-commerce using machine learning

Customer churn is a major challenge faced by e-commerce companies, as it leads to loss of revenue and decreased customer loyalty. In recent years, for predicting and reducing client churn machine learning techniques are powerful tools. This research aims to explore the use of machine learning algorithms for predicting customer churn, annual spending, and product on-time delivery in e-commerce. The study first conducted a comprehensive review of the literature on customer churn in machine learning. The literature showed that customer churn has been predicted successfully using a variety of machine learning algorithms, including support vector machine (SVM), random forest, and decision tree in various industries. To address this gap in the literature, the study conducted an empirical analysis of customer churn in e-commerce using machine learning algorithms. The data were then pre-processed and analyzed utilizing machine learning techniques for prediction. According to the study’s findings, machine learning algorithms are effective in predicting customer churn, and product on-time delivery in e-commerce. The best-performing algorithm SVM achieved an accuracy of 83.45% in predicting customer churn and 68.42% for product on-time delivery prediction.

AB093. Pixel-wise classification of glioma using deep learning for accurate tumour mapping on magnetic resonance imaging.

Central nervous system (CNS) tumours, especially glioma, are a complex disease and many challenges are encountered in their treatment. Artificial intelligence (AI) has made a colossal impact in many walks of life at a low cost. However, this avenue still needs to be explored in healthcare settings, demanding investment of resources towards growth in this area. We aim to develop machine learning (ML) algorithms to facilitate the accurate diagnosis and precise mapping of the brain tumour. We queried the data from 2019 to 2022 and brain magnetic resonance imaging (MRI) of glioma patients were extracted. Images that had both T1-contrast and T2-fluid-attenuated inversion recovery (T2-FLAIR) volume sequences available were included. MRI images were annotated by a team supervised by a neuroradiologist. The extracted MRIs thus obtained were then fed to the preprocessing pipeline to extract brains using SynthStrip. They were further fed to the deep learning-based semantic segmentation pipelines using UNet-based architecture with convolutional neural network (CNN) at its backbone. Subsequently, the algorithm was tested to assess the efficacy in the pixel-wise diagnosis of tumours. In total, 69 samples of low-grade glioma (LGG) were used out of which 62 were used for fine-tuning a pre-trained model trained on brain tumor segmentation (BraTS) 2020 and 7 were used for testing. For the evaluation of the model, the Dice coefficient was used as the metric. The average Dice coefficient on the 7 test samples was 0.94. With the advent of technology, AI continues to modify our lifestyles. It is critical to adapt this technology in healthcare with the aim of improving the provision of patient care. We present our preliminary data for the use of ML algorithms in the diagnosis and segmentation of glioma. The promising result with comparable accuracy highlights the importance of early adaptation of this nascent technology.

Forecasting alzheimer's disease in its early stages with the use of machine learning algorithms

When it comes to dementia in the elderly, Alzheimer's disease (AD) takes the cake. Metabolic disorders impact a huge portion of the global population; as a result, there is a growing interest in using machine learning to better understand these conditions. The annual increase in their incidence rates is quite concerning. Neurodegenerative alterations impact the brain in Alzheimer's disease. Diseases impacting memory and functioning will affect healthcare providers, patients, and families at an increasing rate as the population ages. The monetary, social, and economic spheres will all feel these impacts to a significant degree. It is difficult to foretell the onset of Alzheimer's disease in its early stages. Compared to treatments administered later in the disease's progression, those administered early in the disease's progression are more effective and have less mild side effects. Finding the optimal parameters for Alzheimer's disease prediction has involved the use of several techniques, including Decision Tree, Random Forest, Support Vector Machine, Gradient Boosting, and Voting classifiers. Parameters such as F1-score for ML models, Precision, Recall, and Accuracy are used to evaluate the performance of models that use the Open Access Series of Imaging Studies (OASIS) data to make predictions about Alzheimer's disease. Medical professionals can use the suggested categorization system to identify these illnesses. Reducing yearly mortality rates of Alzheimer's disease through early diagnosis using these ML algorithms is tremendously advantageous. When tested using AD test data, the suggested approach outperforms the competition with an average validation accuracy of more than 90%. In contrast to previous studies, this one achieved a far higher test accuracy score.

Interpretable support vector machine for authentication of omega-3 fish oil supplements using Raman spectroscopy

The use of machine learning algorithms to develop automated analytical methods and smart sensors has drastically increased in recent years. Although these algorithms often provide great performances, the poor interpretability is one of their main limitations. In this work, we present an efficient method to authenticate certified omega-3 fish oil supplements which relies on a vibrational spectroscopic technique (Raman spectroscopy) and a chemically interpretable machine learning algorithm (interval support vector machine, iSVM). The Raman spectra were collected from 248 certified and 520 non-certified omega-3 fish oil samples in a non-destructive fashion. Key differences in the chemical composition of certified and non-certified supplements were unveiled by correlating the iSVM discriminant capability with Raman bands in the corresponding spectral intervals. Proper figures of merit (e.g., accuracy and F1 score) were estimated and critically discussed, and an additional external validation was performed with independent samples to assess the model's robustness and generalization capability. In addition to the excellent chemical interpretability, which revealed the great importance of Raman bands of unsaturated fats to authenticate samples, the proposed Raman-iSVM method showed to be accurate (96 and 92% for internal and external validation sets, respectively), rapid, portable, and computationally cost-effective. Therefore, we believe this method can readily be implemented for in-situ quality monitoring of omega-3 fish oil supplements and other food products.

Forecasting the Value of Real Estate Using Machine Learning Tools

Correct valuation of real estate plays a crucial role in the process of buying and selling. We have carefully studied the existing applications with which we carry out real estate transactions, described their features, advantages and disadvantages. The developed model will help sellers get an estimate of their property according to the parameters entered, which can serve as a starting point for establishing the final value. The computation of real estate values has historically been based primarily on the method of analyzing data manually and subjective estimates, often resulting in errors and delays. The use of machine learning algorithms in solving this problem turned out to be effective, since it has a number of advantages over the manual estimation method, namely: a high level of accuracy, elimination of subjectivity and bias in estimates, time efficiency, cost reduction, use of geospatial data and substantiation of results. The process of creating a machine learning model is conditionally decomposed into four stages, which include collecting data, filtering, processing, supplementing, dividing into different samples and training the model based on this data. We considered the most popular regression algorithms, briefly described the principle of their work, as well as metrics with which you can evaluate the quality of the predicted values of the models. Standard parameters were used to test linear regression algorithms, decision tree, nearest neighbor method, support vector method, and random forest. The determination coefficient R-square is chosen as the main metric. Comparing the coefficient of determination of the results, it became clear that the algorithm “random forest” showed the best result. Having manually selected hyper parameters for this algorithm, the average value of the absolute error of the predicted value is 8.49 %, and the median is 1.9 %. The constructed model meets the established quality requirements and is ready for implementation in the information system of forecasting the value of real estate. For buyers, this service will be relevant, since they will be able to search for real estate according to the parameters entered by them, which have a favorable price for the purchase.

The use of machine learning algorithms to predict financial statement fraud

The use of machine learning algorithms to predict financial statement fraud

Analysis of support vector machine and random forest models for classification of the impact of technostress in covid and post-covid era

This study addresses the growing concern of technostress, a condition caused by the overwhelming use of digital technologies, exacerbated by the COVID-19 pandemic. The researchers developed a predictive model using machine learning algorithms, Random Forest (RF) and Support Vector Machine (SVM), to assess and manage technostress levels. The model considers factors such as age, gender, technology usage hours, and technological experiences to classify stress levels into high, moderate, and low categories. The study collected data through a questionnaire administered to knowledgeable respondents, using a non-probabilistic sampling approach. The results showed that both RF and SVM algorithms achieved high accuracy in classifying technostress, with SVM performing slightly better (94.5% vs 84.50%). The model’s effectiveness in predicting stress levels for users with varying degrees of stress is a significant contribution to the field. The research also developed an interactive user interface to facilitate user engagement with the model, promoting stress management and well-being in a technology-driven society. The study’s findings provide valuable insights into the challenges posed by technostress and offer a solution for mitigating its effects. The use of machine learning algorithms to classify gender based on the dataset demonstrates the model’s potential applications in various areas. Overall, this study demonstrates the importance of addressing technostress in the digital age and provides a valuable tool for managing stress levels. The development of predictive models like this one can help individuals and organizations mitigate the negative impacts of technostress, promoting a healthier and more sustainable relationship with technology.

Violence Detection Using Wi-Fi and 5G/6G Sensing Technologies: A Review

Violence, a pervasive societal concern, demands innovative approaches for its early detection and prevention. This review paper explores the intersection of violence detection and wireless fidelity (Wi-Fi), alongside fifth-generation (5G) and sixth-generation (6G) mobile technologies. Wi-Fi sensing, initially employed for human activity detection, has also demonstrated versatility across a number of other important applications. The significance of leveraging Wi-Fi sensing for violence detection is investigated, underscoring its ability to enhance security protocols and minimise response time. Moreover, through the development and use of machine learning algorithms to analyse and interpret intricate channel state information (CSI) features, the accuracy of violence detection can be improved. Furthermore, this investigation delves into the rapidly developing domain of mobile sensing, examining its contribution to the advancement of violence detection functionalities. The potential convergence of 5G and forthcoming 6G sensing technologies increases the effectiveness of violence detection. Through an analysis of Wi-Fi and mobile sensing technologies, this review paper highlights the transformative capacity that their integration may have on approaches to violence prevention and response.

Chatbot using Machine Learning

A chatbot is a computer program that resembles and interprets spoken or written human conversations, enabling people to engage with digital gadgets in the same way they would with a real person.A chatbot can be as basic as a program that responds to a straightforward question with a single line of text. Using prior data sets, machine learning can be used to develop a chatbot. It forecasts the appropriate response by using the knowledge gained from our earlier data sets. It makes use of machine learning algorithms, such as natural language processing algorithms, for prediction. Conventionally designed chatbots have a number of shortcomings, such as a narrow focus and a lack of customisation.Therefore, in order to improve user experience and elevate our chatbots with effective natural language processing, we apply adaptive personalization approaches. These techniques allow us to modify interactions based on individual preferences and past encounters, which in turn fosters greater engagement and connection.Simple duties like responding to often requested questions can be handled by them, as well as more complicated ones like booking reservations and offering tailored recommendations. Users can readily access them as they can be integrated with various systems and platforms, such messaging apps and social media. Chatbots are becoming a necessary tool for organizations to stay competitive in the market, as the need for digital solutions rises. Keywords- Chatbot, Machine Learning, Prediction, Natural Language Processing, Data sets, Customization, User Experience

MATHEMATICAL MODELING OF THE LEXICAL-SEMANTIC FIELD OF ANDREY SHEPTYTSKYI'S PASTORAL LETTERS USING MACHINE LEARNING ALGORITHMS

This article explores the lexical and semantic field of Andrey Sheptytsky's pastoral letters using machine learning algorithms. To study the lexical and semantic field, such algorithms as Word2Vec, TF-IDF, CBOW, Bag-of-Words and Skip-Gram are used. The texts of this prominent figure are studied through the prism of modern methods of natural language processing, which allows for a more detailed identification of the peculiarities of their vocabulary and semantics. The article discusses the process of creating computer models for text analysis, as well as the use of machine learning algorithms for automatic processing and classification of lexical items in pastoral letters. The results obtained allow us to better understand the specifics of the author's language and mentality, as well as to identify patterns in the use of relevant words and expressions in his work. This work opens up new possibilities for the study of pastoral texts and deepening our understanding of their semantic properties.

Low-cost and scalable detection of sparse informal settlements using machine learning in Gcuwa, Eastern Cape, South Africa

ABSTRACT Informative and scalable cartography plays a pivotal role in curbing urban pollution, waste management, and mitigating environmental damage in the development of informal settlements. The contemporary capabilities of cloud computing facilitate streamlined access to comprehensive data repositories, computational infrastructure, and proficient tools that have rapidly advanced the execution of sprawl mapping procedures. This study tests the performance of four machine-learning algorithms, namely: Gradient Boost, K Nearest Neighbor [KNN], Random Forest [RF], and Support Vector Machine [SVM] with data extracted from cloud computing repositories for delineating informal settlements in Gcuwa, Eastern Cape, South Africa, using low-cost datasets. A systematic approach comprising iterative phases, encompassing data acquisition, the development of a training dataset, modeling, and evaluation was employed. The delineation process involved the extraction of both spectral and textural features from Sentinel-2 imagery. The Random Forest algorithm emerged as the top performer, exhibiting the highest levels of accuracy and F1 score, followed by the gradient boosting, support vector machine, and then the K-nearest neighbor algorithms. Consequently, this innovative use of machine learning algorithms with low-cost datasets and the scalable, resilient approach for detecting informal settlements offers a promising avenue for enhancing urban planning and addressing sustainable development challenges.

Fake Product Detection in Python

Abstract : This paper present a machine learning approach for detecting fake product using python. The study focuses on the use of machine learning algorithm to classify product as genuine or counterfeit based on various features. The dataset used for experimentation is preprocessed, and different machine learning models are trained and evaluated for their performance. The results demonstrate the effectiveness of the proposed approach in detecting fake product with high accuracy and provide insight into future research directions. Index Terms – Machine Learning , Support Vector Machines (SVM) , Natural language processing(NLP).

Cooperative Communication Based Protocols for Underwater Wireless Sensors Networks: A Review.

Underwater wireless sensor networks are gaining popularity since supporting a broad range of applications, both military and civilian. Wireless acoustics is the most widespread technology adopted in underwater networks, the realization of which must face several challenges induced by channel propagation like signal attenuation, multipath and latency. In order to address such issues, the attention of researchers has recently focused on the concept of cooperative communication and networking, borrowed from terrestrial systems and to be conveniently recast in the underwater scenario. In this paper, we present a comprehensive literature review about cooperative underwater wireless sensor networks, investigating how nodes cooperation can be exploited at the different levels of the network protocol stack. Specifically, we review the diversity techniques employable at the physical layer, error and medium access control link layer protocols, and routing strategies defined at the network layer. We also provide numerical results and performance comparisons among the most widespread approaches. Finally, we present the current and future trends in cooperative underwater networks, considering the use of machine learning algorithms to efficiently manage the different aspects of nodes cooperation.

Исследование и разработка подхода для подбора кандидата на вакансию с помощью нейронной сети

A service is described that helps companies find suitable candidates for vacancies. The main functions and advantages of this service are considered, as well as examples of successful use by companies. This service is an innovative tool that allows companies to quickly and efficiently find suitable specialists to fill vacancies. It provides companies with the opportunity to use various search methods, such as searching for resumes, recommendations, social networks and other sources. One of the main advantages of this service is its ability to quickly find candidates who meet the company's requirements. This is achieved through the use of machine learning algorithms that analyze candidate data and determine which ones are most suitable for a particular vacancy. In addition, the service provides companies with the opportunity to pre-select candidates, which saves time and resources for conducting interviews

An Effective Air Pollution Prediction Model Using Machine Learning Algorithms

Air pollution is a major environmental concern globally, with both developed and developing countries facing its impacts. In recent years, citizens and governments have become increasingly concerned about the effects of air pollution on human health and have proposed sustainable development initiatives to address this issue. In a recent study, air pollution data from the years 2020-2022 was collected from a secondary data source. The data set included six key input features, including SO2, PM2.5, CO, PM10, NO2, and O3 values. To analyse this data, various machine learning models were employed, including linear regression, multiple linear regression, KNN, random forest regression, decision tree regression, support vector regression, and artificial neural networks. To ensure the accuracy of the predictions, mean square error and R square were used to measure the absolute error and forecast precision. Additionally, the importance of each input feature in air pollution was investigated, providing valuable insights into the factors that contribute to air pollution levels. Overall, the use of machine learning algorithms in air pollution estimation and prediction has significant potential to improve our understanding of this critical environmental issue and inform effective strategies for addressing it.

The Use of Machine Learning Algorithms in Teaching a Foreign Language, Taking into Account the Individual Characteristics of Students

The Use of Machine Learning Algorithms in Teaching a Foreign Language, Taking into Account the Individual Characteristics of Students

Advanced Deep Learning Methodologies in the Diagnosis of Parkinson's Disease: A Comprehensive Review

This study of the literature explores the field of sophisticated deep techniques for diagnosing Parkinson's illness and severity evaluation. The research looks into the use of machine learning algorithms as potential markers of Parkinson's illness in manual illustrations and speech impairments. These algorithms include XGBoost, Neural Networks with Recurrence, and ensemble models. Numerous feature extraction techniques, model comparison assessments, and preprocessing methodologies are investigated in an effort to improve precision and efficacy of Parkinson's illness diagnosis. In order to improve performance and interpretability of the prototype, the importance of feature engineering, ensemble learning techniques, and interpretability techniques like LIME and SHAP values is emphasized. The study also highlights how crucial it is to integrate several models for real-time monitoring systems to assist in the prompt identification and ongoing tracking of Parkinson's illness. Future possibilities for examine include adding other symptoms including handwriting distortion and olfactory sound loss, investigating new algorithms, and using deep learning approaches to improve system performance. Considerable progress in Parkinson's disease diagnosis and therapy can be made by filling up these research gaps and utilizing deep learning. By addressing these problems via additional study and technical developments, PD can be detected and managed earlier and more effectively, which will ultimately improve patient quality of life. Keywords: Keywords: interpretability, early diagnosis, severity assessment, ensemble learning, XGBoost, deep learning, recurrent neural networks,, machine learning, Parkinson's illness, and feature engineering.

Detection of fake face images using lightweight convolutional neural networks with stacking ensemble learning method.

Images and videos containing fake faces are the most common type of digital manipulation. Such content can lead to negative consequences by spreading false information. The use of machine learning algorithms to produce fake face images has made it challenging to distinguish between genuine and fake content. Face manipulations are categorized into four basic groups: entire face synthesis, face identity manipulation (deepfake), facial attribute manipulation and facial expression manipulation. The study utilized lightweight convolutional neural networks to detect fake face images generated by using entire face synthesis and generative adversarial networks. The dataset used in the training process includes 70,000 real images in the FFHQ dataset and 70,000 fake images produced with StyleGAN2 using the FFHQ dataset. 80% of the dataset was used for training and 20% for testing. Initially, the MobileNet, MobileNetV2, EfficientNetB0, and NASNetMobile convolutional neural networks were trained separately for the training process. In the training, the models were pre-trained on ImageNet and reused with transfer learning. As a result of the first trainings EfficientNetB0 algorithm reached the highest accuracy of 93.64%. The EfficientNetB0 algorithm was revised to increase its accuracy rate by adding two dense layers (256 neurons) with ReLU activation, two dropout layers, one flattening layer, one dense layer (128 neurons) with ReLU activation function, and a softmax activation function used for the classification dense layer with two nodes. As a result of this process accuracy rate of 95.48% was achieved with EfficientNetB0 algorithm. Finally, the model that achieved 95.48% accuracy was used to train MobileNet and MobileNetV2 models together using the stacking ensemble learning method, resulting in the highest accuracy rate of 96.44%.

Bibliometric Analysis and Visualization of Scientific Literature on Heart Disease Classification Using a Logistic Regression Model.

With the advancement in artificial intelligence, the use of machine learning algorithms for clinical prediction has increased tremendously. Logistic regression is one of the powerful machine learning algorithms that can be used to predict the probability of a variable. Logistic regression is very popular among medical researchers owing to its simplicity, interpretability, and solid statistical foundation. This study aims to investigate the research productivity of heart disease classification using a logistic regression model to analyze the current patterns and potential future trends through bibliometric analysis. Additionally, it aims to highlight the impact and quality of research in the area, identify prominent research groups,the countries actively contributing to the field, which will help the researchers andhealthcare professionals to pinpoint research gaps, influential authors, and make informed decisions and invest resources accordingly. The data is collected from a database of Scopus spanning from 2019 to 2023. We have used two bibliometric software, Biblioshiny(Aria and Cuccurullo, 2017) and VOSviewer (Centre for Science and Technology Studies (CWTS), Leiden University,the Netherlands), to analyze the bibliographic data regarding the citation count, contribution of authors, publication count, the contribution of institutions, etc.There are 2331 documents under study which were fed into both software to analyze the data. With 700 documents, China topped the list of most productive countries indicating the vast contribution of the country followed by India and the United States. Contributions of the Harvard Medical School, Boston, MA, United States are found to be the greatest with six papers. The most productive author is Wang Y with 73 documents. Analysis of trending topics reveals that the field progressing towards using support vector machines (SVM), k-nearest neighbours (KNN), and naïveBayes algorithms. The article has only considered data from Scopus excluding literature indexed in other databases which limits the potential coverage of the data. Also, the work focuses on recent developments excluding older literature from 2019 which could be a limitation. Furthermore, since the study is a bibliometric analysis targeting the use of logistic regression for heart disease prediction, powerfultechniques such as SVM, decision trees, random forests, neural networksand deep learning have not been included, which could be another limitation.

Filter Bubbles and the Unfeeling: How AI for Social Media Can Foster Extremism and Polarization

Social media have undoubtedly changed our ways of living. Their presence concerns an increasing number of users (over 4,74 billion) and pervasively expands in the most diverse areas of human life. Marketing, education, news, data, and sociality are just a few of the many areas in which social media play now a central role. Recently, some attention toward the link between social media and political participation has emerged. Works in the field of artificial intelligence have already pointed out that there is a close link between the use of machine learning algorithms in social media and possible epistemic isolation, which could lead to political radicalization. The idea supporting this paper is that artificial intelligence for social media can actively put users’ deliberative capacity at risk and foster political extremism. To prove these claims, I proceed along two lines of inquiry. First, I focus on filter bubbles, namely the result of selections made by algorithms that recommend contents that meet users’ expectations and opinions. To analyze this phenomenon, I refer to the Deweyan model of experience. Second, I connect the filter bubbles problem to the Deweyan idea of deliberative and participatory democracy and Nussbaum’s concept of political compassion. The purpose of this paper is to provide a philosophical foundation that can both (1) effectively serve as a method for analyzing machine learning algorithms and their potential problems in relation to political extremism, and (2) be adopted as a standard to counter the danger of extremism associated with social media experience.