Öz Hava kirliliği, günümüzün en büyük sorunlarından birini teşkil etmektedir. Hava kirliliği, nüfusun artması, kentsel gelişme ve büyüme, endüstrinin gelişmesiyle giderek artan bir önem arz etmektedir. Genellikle hava kirleticilerinin insanlara, canlılara ve çevreye zararlı etkileri zaman, mekan, etki süresi, konsantrasyon ve diğer karakteristiklerine bağlı olarak karmaşık dağılım şekilleri göstermektedir. Bu karmaşıklık, kirletici örnekleri ve eğilimleri modelleme veya ölçmede, ayrıca insanların maruz kaldığı seviyeleri tahmin etmenin zor olduğu anlamına gelmektedir. Hava kirliliğini önleme çalışmaları arasında en önemli adımlardan biri hava kirlenmesi olayının bir model içerisinde değerlendirilmesidir. Bu çalışmada Kastamonu ili ele alınarak, meteoroloji ve çevre uygulamalarında oldukça yeni ve başarılı sonuçlar elde edilen çeşitli makine öğrenmesi algoritmaları ile hava kirliliğinin tahmininde, bazı meteorolojik değişkenler kullanılarak hava kirliliği tahmini yapacak modeller geliştirilmiştir. Minimum-Maksimum (Min-Max) normalizasyon tekniği, öğrenme yöntemleri ile birlikte kullanılmıştır. Tahmin modellerinde, Yapay Sinir Ağları (YSA), Rastgele Orman (Random Forest), K-En Yakın Komşu (K-Nearest Neighborhood), Lojistik Regresyon (Logistic Regression), Karar Ağacı (Decision Tree), Lineer Regresyon (Linear Regression) ve Basit Bayes (Naive Bayes) yöntemleri kullanılmıştır. Çalışmada elde edilen performans değerleri, literatürdeki benzer çalışmalarla kıyaslanarak problemin çözümüne ilişkin en uygun tahmin algoritması tespit edilmiştir. Veri setinin %70'i eğitim ve %30'si test verisi olarak ayrılmıştır. Çalışma sonucunda, YSA modeli için doğru tahmin oranı %87 ve diğer makine öğrenmesi modellerinden Rastgele Orman doğruluk oranı %99 ve Karar Ağacı doğruluk oranı %99 değerleri ile tahminlemede en başarılı sonuçları verdiği görülmüştür. Lineer Regresyon yöntemi %30'lık doğruluk oranı ile oldukça kötü performans sergilemektedir. KastamonuDataSet üzerinde kullanılan yöntemlerin performans değerlendirmelerinde Açıklayıcılık Katsayısı (R 2), Ortalama Karesel Hata (Mean Squared Error-MSE), Ortalama Hata Kare Kökü (Root Mean Square Error-RMSE) ve Ortalama Mutlak Hata (Mean Absolute Error-MAE) metrikleri bakımından istatistiksel önemli farklılıkların bulunduğu tespit edilmiştir.
Coronary Artery Disease (CAD) takes place in the category of fatal diseases resulting in death in our country and around the world. Each year about 340 thousand patients lost their lives due to CAD in Turkey. Early diagnosis is essential to reduce risk and prolong lifetime of these patients for diseases that require long-term treatment having death risk like CAD. For this reason, classification of CAD by using medical data processing and machine learning algorithms are important in order to develop assistive or expert systems for physicians. In this study, five different machine learning algorithms were applied to estimate whether patients in the Z-Alizadeh Sani data set extracted from the UCI machine learning pool are CAD. Accuracy, precision, recall, specificity and F1 score were compared as classification performance indicators to evaluate decision tree, random forest (RF), support vector machines (SVM), nearest neighborhood (k-NN) and multi-layer sensor (MLP) methods. According to the evaluation results, the MLP method gave high classification accuracy with 90%. It also appears that RF performs relatively better than other metrics. This results, show that these classification algorithms can be use for helping healthcare systems.
Although Artificial Neural Networks have been used for many years, its use in air pollution analysis has become widespread in 10 years. In this study, an Artificial Neural Network model was proposed to predication air pollution in Kastamonu province of Turkey. In this study as an example of Kastamonu province, Artificial Neural Network model was formed by using a pollution parameter (PM10) and 5 different meteorological factors (air temperature, air pressure, humidity, wind direction and wind speed) which were measured daily data during (2015-2018) period. It is aimed to propose refined model to predict the value of air pollution concentration (SO 2) after 24 hours by using this model. In other words, the air quality model is predicted. Artificial Neural Network is very successful compared to the new and classical statistical methods. Feed back-propagation algorithm has been used in all developed Artificial Neural Network model. The data set used in this study is divided into three subsets including training, validations and test data sets. The first 70% percent of the data set were used as the training subset, 15% of the data set used as the test set and 15% of the data set were used in validation set. The Mean Squeare Error (MSE) was measured for the performance of the network. It was observed that the developed ANN model was in agreement with the experimental results.
Nowadays, there exists a lot of information that can be handled from business transactions and scientific data and information retrieval is simply no longer enough for decision-making. In this paper will supervised machine learning technique is applied to the mine data warehouse for Enterprise Resource Planning (ERP) of the General Electricity Company of Libya (GECOL). This technique has been applied for the first time on the data of production, transportation and distribution departments. These data are in the form of purchase and work orders of operational material strategic equipment spare parts. This technique would extract prediction rules in order to assist the decisionmakers of the company to make appropriate future decisions more easily and in less time. A supervised machine learning technique has been adopted and applied for the mining data warehouse. A well-known software package for data mining which is referred to as WEKA tool was adopted throughout this work. The WEKA tool is applied to the collected data from GECOL. The conducted experiments produce prediction models in the form set of rules in order to help responsible employees make the suitable, right and accurate future decision in a simple way and inappropriate time. The collected data were preprocessed to be prepared in a suitable format to be fed to the WEKA system. A set of experiments has been conducted on those data to obtain prediction models. These models are in the form of decision rules. The produced models were evaluated in terms of accuracy and production time. It can be concluded that the obtained results are very promising and encouraging.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
334 Leonard St
Brooklyn, NY 11211
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.