logistic regression
در نشریات گروه برق-
Scientia Iranica, Volume:31 Issue: 22, Nov-Dec 2024, PP 2139 -2147An insufficient amount of capital conservation buffer would cause a financial institution to be unable to withstand fluctuations in the economic cycle; while an excessive amount would reduce the financial institution’s available funds, which would lead to a loss of the capital available for investment. In order to address this issue in an effective manner, the business loan default prediction model is established in this study by integrating survival analysis with logistic regression. In the section of case validation, the reliability of the proposed approach is validated with the information of businesses that have been granted loans by financial institutions in Taiwan, and the proposed approach was also compared with the Cox proportional hazards model approach, which is frequently applied by financial institutions. The empirical results demonstrate that the approach proposed in this study could predict a business loan default state closer to the actual default trend, and provide prediction results superior to that of the Cox proportional hazards model, thus, providing financial institutions with effective and reliable information for reference, which will allow them to prepare an appropriate amount of capital conservation buffer, and improve the capital flexibility of the financial institution.Keywords: Business Loan Default Prediction, Small, Medium-Sized Enterprises, Basel Capital Accord, Survival Analysis, Logistic Regression
-
نشریه پژوهش های نظری و کاربردی هوش ماشینی، سال دوم شماره 1 (پیاپی 3، بهار و تابستان 1403)، صص 104 -119یکی از وظایف سازمان های آتش نشانی ارائه گزارش حریق و حوادث به مراجع قضائی، بیمه و سایر نهادهای درخواست کننده جهت تصمیم گیری و پرداخت خسارت است، لذا یافتن ماهیت حریق به نحوی که مولفه های غیرعملیاتی در تصمیم کارشناسان آتش نشانی کمترین تاثیر را داشته باشد، اهمیت این پژوهش را بیشتر خواهد نمود. باتوجه به اینکه حدود 1 درصد از گزارشات حریق این سازمان ماهیت نامعلوم دارند، این موضوع باعث سردرگمی در ارائه خدمات مناسب به ارباب رجوع را داشته و تصمیم گیری با مشکل مواجه شده است. هدف از این پژوهش پیش بینی ماهیت حریق مبتنی بر الگوریتم های یادگیری ماشین در شهر مشهد می باشد. در این پژوهش ابتدا مجموعه داده حریق 7 ساله (1395-1401) مورد بررسی و واکاوی قرار گرفت و پس از آن با توجه به مسئله و ادبیات موضوع و با انجام پیش پردازش و مهندسی ویژگی مجموعه داده ای با تعداد 46 ویژگی و 28930 نمونه تهیه شد. در مرحله بعد برای پیش بینی ماهیت حریق از سه الگوریتم یادگیری ماشین با ناظر استفاده شد و نتایج آن ها با هم مقایسه شد که الگوریتم رگرسیون لجستیک با 79.66 درصد دقت با زمان اجرای 1 ثانیه نتیجه بهتری را بین سه الگوریتم جهت پیش بینی ماهیت حریق ایجاد نموده است.کلید واژگان: ماهیت حریق، عمدی، غیرعمدی، یادگیری ماشین، رگرسیون لجستیکOne of the duties of firefighting organizations is to submit fire and accident reports to judicial authorities, insurance, and other requesting institutions to make decisions and pay damages, therefore, finding the nature of the fire in such a way that non-operational components have the least impact on the decision of firefighting experts increases the importance of this research. Considering that about 1% of the fire reports of this organization are unknown, this issue has confused providing proper services to Arbab-Rojoo and decision-making has been difficult. This research aims to predict the nature of fire based on machine learning algorithms in the city of Mashhad. In this research, the 7-year fire data set (1395-1401) was first examined and analyzed, and then, according to the problem and the literature, a data set with 46 features and 28930 samples was prepared by pre-processing and feature engineering. In the next step, to predict the nature of the fire, three machine learning algorithms were used with the observer and their results were compared. The logistic regression algorithm, with 79.66% accuracy and an execution time of 1 second, created a better result between the three algorithms in predicting the nature of the fire.Keywords: Nature Of Fire, Intentional, Unintentional, Machine Learning, Logistic Regression
-
The incidence of heart-related illnesses is on the rise worldwide. Heart diseases are primarily caused by a multitude of parameters, including high blood pressure, diabetes, and excessive cholesterol, which are controlled by poor dietary and lifestyle choices. The growth in cardiovascular diseases (CVD) is mostly due to several other behaviors, such as smoking, drinking, and sleeplessness. In the research, machine learning-based prediction methods work on the audio recordings of heartbeats known as phonocardiograms (PCG) to develop an algorithm that differentiates a normal healthy heart from an abnormal heart based on the heart sounds. The data set consists of 831 normal and 260 abnormal data, and the duration of each sample is 5 seconds. Features extracted from the data are up-sampled and applied to the logistic regression and random forest classification models. The developed models record a classification accuracy of 71% for logistic regression and 94% for the random forest model. Further, artificial neural networks (ANN) and Deep learning networks have been trained to improve performance and demonstrated an accuracy of 94.5%.
Keywords: Phonocardiogram (PCG), Machine Learning, Logistic Regression, Random Forest, Deep Learning -
با توجه به افزایش روزافزون داده، حجم داده های متنی نیز با سرعت بالایی در حال رشد است. استخراج اطلاعات از این داده های متنی یکی از ضرورت های دنیای مبتنی بر اطلاعات امروزی است. دسته بندی متن یکی روش های دست یابی به اطلاعات این داده های حجیم است. در این تحقیق با استفاده از یک مجموعه داده استاندارد اخبار فارسی که شامل پنج ویژگی در بیش از 86هزار خبر بود به بررسی عملکرد الگوریتم رگرسیون لجستیک در دسته بندی متن فارسی و همچنین مقایسه آن با سایر کارهای مشابه پرداختیم. با توجه مراحل ساخت یک دسته بند متن،روش مورد استفاده در بخش بردارسازی را توضیح داده و همچنین اهمیت بخش پیش پردازش و مخصوصا روش مورد استفاده در برچسب گذاری و تبدیل برچسب های فرعی به اصلی را بیان کردیم. در ارزیابی نهایی، با استفاده از تغییر پارامترهای الگوریتم و همچنین اصلاح برچسب های اخبار، به نتیجه مطلوب 95% در معیار دقت برای دسته بندی متن مجموعه داده اخبار فارسی رسیدیم.
کلید واژگان: دسته بندی متن، رگرسیون لجستیک، پیش پردازش متن، مجموعه داده اخبار فارسیJournal of New Achievements in Electrical, Computer and Technology, Volume:3 Issue: 6, 2023, PP 24 -37Due to the ever-increasing amount of data, the amount of textual data is also growing at a high speed. Extracting information from these textual data is one of the necessities of today's information-based world. Text classification is one of the methods of obtaining information from this massive data. In this research, using a standard dataset of Persian news, which included five features in more than 86 thousand news, we investigated the performance of the logistic regression algorithm in the classification of Persian text and also compared it with other similar works. Considering the steps of creating a text category, we have explained the method used in the vectorization section and also stated the importance of the pre-processing section, especially the method used in tagging and converting sub-tags to main ones. In the final evaluation, by changing the algorithm's parameters and modifying the news tags, we reached the desired result of 95% in the accuracy criterion for the text classification of the Persian news dataset.
Keywords: Text classification, Logistic regression, Text preprocessing, Persian news dataset -
امروزه، رویکردهای یادگیری ماشین نقش مهمی را در شناسایی عوامل موثر بر کیفیت محصولات تولیدی ازجمله تولید کاشی و سرامیک ایفا می کنند. یکی از چالش های موجود در تولید کاشی و سرامیک، معیوب شدن کاشی ها به دلیل ایجاد انحراف در ابعاد کاشی تولیدی است. در صورتی که بتوان با توجه به پارامترهای فرآیند تولید، امکان ایجاد انحراف در ابعاد کاشی را قبل از تولید پیش بینی نمود، می توان از تولید کاشی معیوب جلوگیری و نسبت به تنظیم مجدد پارامترهای تولید اقدام نمود. در این پژوهش، یک سیستم خودکار جهت پیش بینی دسته ی انحراف کاشی و شناسایی عوامل تاثیرگذار بر آن، پیشنهاد شده است. بدین منظور سه طبقه بندکننده ی مختلف شامل رگرسیون منطقی، جنگل تصادفی و ماشین بردار پشتیبان جهت مدل سازی پارامترهای مربوطه مورد بررسی قرارگرفته و برترین ساختار معرفی شده است. علاوه بر این با بررسی چند دسته ویژگی و بهره گیری از روش انتخاب ویژگی پیش رو، متغیرهای موثر در تصمیم گیری انحراف کاشی نیز شناسایی شده اند. نتایج آزمایش های انجام شده بر روی نمونه های واقعی، نشان می دهد که رویکرد جنگل تصادفی کارایی بهتری نسبت به رویکردهای دیگر داشته و تاثیرگذارترین پارامترها در ایجاد انحراف کاشی، مقدار نامناسب دماهای کوره بوده است.
کلید واژگان: دسته بندی خودکار، انحراف ابعاد کاشی، یادگیری ماشین، جنگل تصادفی، ماشین بردار پشتیبان، رگرسیون منطقی، انتخاب ویژگی پیش روJournal of Iranian Association of Electrical and Electronics Engineers, Volume:19 Issue: 2, 2022, PP 199 -206In recent years, machine learning approaches play an important role in quality identification of manufactured products including tiles and ceramics. Deviation of tile dimensions is one the main challenge in ceramic and tile industry. Prediction of this deformation will be beneficial if it can be predicted before producing the tile. In this paper, an automatic system has been proposed to predict the deviation of the ceramic tiles. Besides, a machine learning approach is utilized to identify the most effective parameters that leads to tiles’ defect. In this way, three different classification approaches including logistic regression, random forest, and support vector machine have been studied and the best solution is determined for this purpose. Moreover, several feature sets and forward feature selection method have been employed to select more effective variables on our decision making. The experimental results conducted on real-world dataset show that, random forest approach achieves better performance than others, and the results illustrate that improper temperature parameters has more effect on tile deviation.
Keywords: Automatic classification, Dimensional deviation of tile, Machine learning, Random forest, Support vector machine, Logistic regression, Forward feature selection -
The use of accurate lithological maps is inevitable in the preparation of rock unit’s erosion susceptibility maps. In this study, rock unit outcrops in the Soh Basin (50 km Northern Isfahan) were extracted using nonlinear correlation analysis of satellite data. Moreover, rock unit’s erosion susceptibility such as marl, shale, and quaternary deposits and resistant rock units such as sandstone and limestone were extracted based on soil erosion intensity factors. The lithology of the basin was studied using the virtual variables method. Initially, rock units, as a virtual independent variable, and the PC1 (the first principal component) of ETM+ multispectral bands were analyzed by a multiple linear regression model. Afterward, rock units were analyzed in logistic regression analysis as virtual dependent variables. The results revealed that logistic regression analysis is a suitable model for rock unit’s extraction.
Keywords: Satellite Data, Landsat ETM +, Lithological Mapping, Soil Erosion Susceptibility, Logistic Regression -
the lithological maps is inevitable in the preparation of rock unit’s erosion susceptibility maps. In this study, rock unit outcrops in the Soh Basin (50 km Northern Isfahan) were extracted using nonlinear correlation analysis of satellite data. Moreover, rock unit’s erosion susceptibility such as marl, shale, and quaternary deposits and resistant rock units such as sandstone and limestone were extracted based on soil erosion intensity factors. The lithology of the basin was studied usingthe virtual variables method. Initially, rock units, as a virtual independent variable, and the PC1 (the first principal component) of ETM+ multispectral bands were by amultiple linear regression model. Afterward, rock units were in logistic regression analysis as virtual dependent variables. The results revealed that logistic regression analysis is a suitable model for rock unit’s extraction.Keywords: satellite data, Landsat ETM +, lithological mapping, soil erosion susceptibility, Logistic regression
-
Due to the increasing development and applications of the Internet of Things (IoT), detection and prevention of intruders into the network and devices has gained much attention in the past decade. For this challenge, traditional solutions of Intrusion Detection Systems (IDS) are not responsive in IoT environments or at least may not be very efficient. In this article, we deeply investigate the previous methods of using machine learning methods for intrusion detection in IoT, and two methods for feature extraction and classification are proposed. The first method is feature extraction and classification using Logistic Regression (LR) and the second method is to use an Artificial Neural Network (ANN) for classification. To evaluate the performance of the proposed method, six devices of the N_BaIoT dataset, which consists of data samples related to nine devices IoT and several attacks are used according to some criteria for evaluating the performance of the proposed methods. Simulation results in comparison with some other deep learning methods in terms of accuracy, precision, recall and F1-score show that using logistic regression, is more efficient and above 90% classification accuracy is achieved.Keywords: Internet of Thing, Anomaly Detection, artificial neural network, Logistic Regression, Botnet
-
The earth surface is itself a complex system, and land cover variation is a complex process influenced by the interference of variables. In this study, the data of Sentinel 2 for 2017 and 2016 were processed and classified to study the changes in the Andika area. After discovering vegetation changes between two images over the mentioned time, vegetation increased by 661.74 hectares. Multiple regressions have been used to identify factors affecting vegetation changes. Multiple regressions can explain the relationship between vegetation changes and the factors affecting them. In order to investigate the factors affecting vegetation change, altitude data, distance from the road, distance from residential areas of the village and river were introduced into regression equation. Since this method uses three parameters such as Pseudo-R2 and Relative Operation Characteristic (ROC(, 0.23, and 0.696 values for the above parameters, which indicates that the model is in good agreement. The results of regression analysis show that linear composition of height variable as independent variables in comparison with other parameters has been able to estimate vegetation change. Subsequently, by using two classified pictures of 2017 and 2016, the amount of vegetation changes was calculated, and Markov chain method was used for 2018 forecast changes.Keywords: NDVI, Sentinel 2, Cellular Automata Markov, logistic regression
-
این مقاله مسئله طبقه بندی شورایی را برای تصدیق گوینده مستقل از متن بررسی می کند. ازآنجاکه ممکن است یک طبقه بند از اطلاعات مختلف سیگنال گفتار بهره نبرد، استفاده از یک طبقه بند برای تصدیق گوینده ممکن است منجر به تصمیم قابل اعتمادی نشود. بنابراین بهترین سامانه های تصدیق گوینده از مجموعه ای از طبقه بندهای مکمل برای رسیدن به تصمیمات قابل اعتماد استفاده می کنند. در اکثر مطالعات اخیر که روی ترکیب طبقه بندها برای تصدیق گوینده انجام شده است، ترکیب خطی وزن داری از امتیاز طبقه بندهای خبره پایه برای رسیدن به امتیاز نهایی تصدیق استفاده می شود که وزن های این ترکیب با استفاده از روشی مانند رگرسیون لجستیک و در زمان آموزش به دست می آیند. در این تحقیقات مسائلی از قبیل همبستگی بین طبقه بندها و برتری برخی طبقه بندها برای برخی داده آزمون به خوبی در نظر گرفته نشده است. در این مقاله با استفاده از فرایند طراحی شورا و قاعده ترکیب بر اساس داده آزمون برای هر دو مسئله راه حلی ارائه می شود. بررسی های انجام شده ما روی دادگان ارزیابی تصدیق گوینده NIST 2004 نشان می دهد روش پیشنهادی در مقایسه با روش مبنای ترکیب تنک طبقه بندها کارایی مناسبی را دارد.کلید واژگان: بازشناسی گوینده، تصدیق گوینده، طبقه بندی شورایی، طبقه بندی شورایی تطبیقی، رگرسیون لجستیکThis paper considers the ensemble classification for the text independent speaker verification issue. Using one classifier for the speaker verification may not result in dependable decision, because it may not exploit different characteristics of speech signal. Therefore, state-of-the-art speaker verification systems use an ensemble of classifiers for the verification. Most of the ensemble speaker verification systems use a weighted summation of the score of the individual expert classifiers to calculate the final score of the verification. The weights of this score fusion is obtained using a method, e.g. logistic regression, in the training phase. These works do not efficiently take into account issues such as correlation of classifiers and instance specific behavior of the base classifiers into account. In this paper a new solution is proposed for these two issues by using the process of ensemble design and combination rule based on training data. The obtained results on NIST 2004 speaker evaluation corpus show the effectiveness of the proposed methods in comparison to the sparse classifier fusion, as a baseline method.Keywords: Speaker recognition, speaker verification, ensemble classification, adaptive ensemble classification, logistic regression
- نتایج بر اساس تاریخ انتشار مرتب شده‌اند.
- کلیدواژه مورد نظر شما تنها در فیلد کلیدواژگان مقالات جستجو شده‌است. به منظور حذف نتایج غیر مرتبط، جستجو تنها در مقالات مجلاتی انجام شده که با مجله ماخذ هم موضوع هستند.
- در صورتی که می‌خواهید جستجو را در همه موضوعات و با شرایط دیگر تکرار کنید به صفحه جستجوی پیشرفته مجلات مراجعه کنید.
©2000-2025 magiran.com All Rights Reserved.