11 sonuçlar
Arama Sonuçları
Listeleniyor 1 - 10 / 11
Yayın Reviewing the effects of spatial features on price prediction for real estate market: Istanbul case(IEEE, 2022-09-16) Ecevit, Mert İlhan; Erdem, Zeki; Dağ, HasanIn the real estate market, spatial features play a crucial role in determining property appraisals and prices. When spatial features are considered, classification techniques have been rarely studied compared to regression, which is commonly used for price prediction. This study reviews spatial features' effects on predicting the house price ranges for real estate in Istanbul, Turkey, in the classification context. Spatial features are generated and extracted by geocoding the address information from the original data set. This geocoding and feature extraction is another challenge in this research. The experiments compare the performance of Decision Trees (DT), Random Forests (RF), and Logistic Regression (LR) classifier models on the data set with and without spatial features. The prediction models are evaluated based on classification metrics such as accuracy, precision, recall, and F1-Score. We additionally examine the ROC curve of each classifier. The test results show that the RF model outperforms the DT and LR models. It is observed that spatial features, when incorporated with non-spatial features, significantly improve the prediction performance of the models for the house price ranges. It is considered that the results can contribute to making decisions more accurately for the appraisal in the real estate industry.Yayın Hotel sales forecasting with LSTM and N-BEATS(IEEE, 2023-09-15) Özçelik, Şuayb Talha; Tek, Faik Boray; Şekerci, ErdalTime series forecasting aims to model the change in data points over time. It is applicable in many areas, such as energy consumption, solid waste generation, economic indicators (inflation, currency), global warming (heat, water level), and hotel sales forecasting. This paper focuses on hotel sales forecasting with machine learning and deep learning solutions. A simple forecast solution is to repeat the last observation (Naive method) or the average of the past observations (Average method). More sophisticated solutions have been developed over the years, such as machine learning methods that have linear (Linear Regression, ARIMA) and nonlinear (Polynomial Regression and Support Vector Regression) methods. Different kinds of neural networks are developed and used in time series forecasting problems, and two of the successful ones are Recurrent Neural Networks and N-BEATS. This paper presents a forecasting analysis of hotel sales from Türkiye and Cyprus. We showed that N-BEATS is a solid choice against LSTM, especially in long sequences. Moreover, N-BEATS has slightly better inference time results in long sequences, but LSTM is faster in short sequences.Yayın Predictive vector quantization of 3-D polygonal mesh geometry by representation of vertices in local coordinate systems(IEEE, 2005) Bayazıt, Uluğ; Orcay, Özgür; Konur, Umut; Gürgen, Sadık FikretA large family of lossy 3-D mesh geometry compression schemes operate by predicting the position of each vertex from the coded neighboring vertices and encoding the prediction error vectors. In this work, we first employ entropy constrained extensions of the predictive vector quantization and asymptotically closed loop predictive vector quantization techniques that have been suggested in [3] for coding these prediction error vectors. Then we propose the representation of the prediction error vectors in a local coordinate system with an axis coinciding with the surface normal vector in order to cluster the prediction error vectors around a 2-D subspace. We adopt a least squares approach to estimate the surface normal vector from the non-coplanar, previously coded neighboring vertices. Our simulation results demonstrate that the prediction error vectors can be more efficiently vector quantized by representation in local coordinate systems than in global coordinate systems.Yayın Graph convolutional network based virus-human protein-protein interaction prediction for novel viruses(Elsevier Ltd, 2022-08-13) Koca, Mehmet Burak; Nourani, Esmaeil; Abbasoğlu, Ferda; Karadeniz, İlknur; Sevilgen, Fatih ErdoğanComputational identification of human-virus protein-protein interactions (PHIs) is a worthwhile step towards understanding infection mechanisms. Analysis of the PHI networks is important for the determination of path-ogenic diseases. Prediction of these interactions is a popular problem since experimental detection of PHIs is both time-consuming and expensive. The available methods use biological features like amino acid sequences, molecular structure, or biological activities for prediction. Recent studies show that the topological properties of proteins in protein-protein interaction (PPI) networks increase the performance of the predictions. The basic network projections, random-walk-based models, or graph neural networks are used for generating topologically enriched (hybrid) protein embeddings. In this study, we propose a three-stage machine learning pipeline that generates and uses hybrid embeddings for PHI prediction. In the first stage, numerical features are extracted from the amino acid sequences using the Doc2Vec and Byte Pair Encoding method. The amino acid embeddings are used as node features while training a modified GraphSAGE model, which is an improved version of the graph convolutional network. Lastly, the hybrid protein embeddings are used for training a binary interaction classifier model that predicts whether there is an interaction between the given two proteins or not. The proposed method is evaluated with comprehensive experiments to test its functionality and compare it with the state-of-art methods. The experimental results on the benchmark dataset prove the efficiency of the proposed model by having a 3–23% better area under curve (AUC) score than its competitors.Yayın A novel regression method for software defect prediction with kernel methods(2013) Okutan, Ahmet; Yıldız, Olcay TanerIn this paper, we propose a novel method based on SVM to predict the number of defects in the files or classes of a software system. To model the relationship between source code similarity and defectiveness, we use SVM with a precomputed kernel matrix. Each value in the kernel matrix shows how much similarity exists between the files or classes of the software system tested. The experiments on 10 Promise datasets indicate that SVM with a precomputed kernel performs as good as the SVM with the usual linear or RBF kernels in terms of the root mean square error (RMSE). The method proposed is also comparable with other regression methods like linear regression and IBK. The results of this study suggest that source code similarity is a good means of predicting the number of defects in software modules. Based on the results of our analysis, the developers can focus on more defective modules rather than on less or non defective ones during testing activities.Yayın Investigating effects of milling conditions on cutting temperatures through analytical and experimental methods(Elsevier Science SA, 2018-12) Karagüzel, Umut; Budak, ErhanCutting temperatures in milling operations have a significant impact on tool wear, size and shape tolerances and residual stresses of the machined part. Prediction and measurement of cutting temperatures in milling, on the other hand, have some challenges due to the rotary tools resulting in an intermittent process and transient thermal loadings. In this study, novel approaches are presented to model and measure the cutting tool temperature variations during milling. The model is used to predict effects of milling conditions on cutting temperatures particularly to determine a relationship between tool temperature and radial depth of cut. The model predictions are verified by measurements obtained from the developed measurement technique and the literature data.Yayın Forecasting and analysis of energy consumption and waste generation in Antalya with SVR(IEEE, 2023-12-24) Özçelik, Şuayb Talha; Tek, Faik Boray; Şekerci, ErdalAntalya, a rapidly expanding coastal city in Türkiye, has experienced significant changes due to urbanization and increasing tourism activities. Comprehending tourism trends is crucial for the city's sustainable development and environmental management. Based on this perspective, this paper aims to present a comprehensive retrospective analysis of Antalya's energy consumption, domestic solid waste generation, wastewater generation, population growth, and tourist numbers over the years. Antalya faces significant challenges due to escalating trends in listed areas. Utilizing the Support Vector Regression, this study projects a need for an additional 1715 GWh of electricity production capacity, an expansion of wastewater capacity by 85639 thousand m3, and an increase in domestic solid waste disposal capacity by 597745 tons by 2028 to accommodate growing demands. We emphasize the importance of adopting effective policies and strategies to support energy efficiency, waste reduction, and wastewater management alongside sustainable urban planning and tourism management for Antalya's long-Term environmental sustainability and development. The findings presented in this study provide valuable insights for policymakers, urban planners, and stakeholders to make informed decisions, ensuring a balanced approach toward economic growth and environmental conservation.Yayın Transient multi-domain thermal modeling of interrupted cutting with coated tools(Springer Science and Business Media Deutschland GmbH, 2021-09) Karagüzel, UmutInterrupted cutting operations, such as milling, produce fluctuating tool temperatures which directly affect the process outputs. Thus, prediction of cutting tool temperatures enables process planning, selection of materials for tool substrate and coating layers, and tool geometric design for improved productivity in machining operations. Theoretical analysis of temperature is a cost effective way to predict the tool temperatures. Considering the industrial needs, a theoretical model should be fast, easy to implement, and reliable. To that end, a novel hybrid model, which assembles analytical and numerical methods, is proposed in this study. This novel transient thermal model simulates the interrupted cutting with coated cutting tools. The proposed model includes an analytical heat flux calculation at the tool-chip interface considering the sticking-sliding contact behavior. The determined heat flux is, then, used to perform a numerical solution of the transient heat conduction problem in the cutting tool geometry with temperature-dependent thermal properties. The developed model is validated with experimental results found in literature under different cutting conditions. The results show that the model can predict the maximum temperatures generated in a thermal cycle with an accuracy of 2–10%. Thus, the proposed model can be further used to determine the process parameters, properties of coating layers, and tool geometric design.Yayın Backcasting Bitcoin volatility: ARCH and GARCH approaches(Suat Teker, 2024-12-31) Teker, Dilek; Teker, Suat; Demirel Gümüştepe, EsinPurpose- The primary purpose of this study is to model Bitcoin price volatility and forecast its future price returns using advanced econometric models such as ARCH and GARCH. The study aims to enhance risk management strategies and support informed investment decisions by addressing the time-varying nature of Bitcoin’s volatility. The research explores the persistence of volatility shocks and the clustering of price movements to provide insights into market dynamics. Methodology- This research examines daily Bitcoin closing prices over the period from January 2020 to October 2024. The data was preprocessed to ensure reliability, including applying logarithmic transformations to standardize the data and eliminate trends. Stationarity tests, such as the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and KPSS tests, were conducted to confirm the series' stationarity. The ARCH-LM test was utilized to detect volatility clustering which is essential for validating the use of ARCH and GARCH models. Following this, ARIMA models were employed to define mean equations and GARCH models were used to estimate conditional variance and capture volatility dynamics. The dataset was split into training and validation subsets with data from July to October 2024 reserved for validation. Findings- The findings demonstrate that Bitcoin’s price movements exhibit significant volatility clustering and persistence of shocks which are key characteristics effectively captured by ARCH and GARCH models. These models provide valuable insights into the volatility patterns of Bitcoin, supporting their application in cryptocurrency analysis. Despite their robustness, the models face limitations in precise return forecasting during highly volatile periods, suggesting the need for further refinement or integration with advanced approaches. Conclusion- The research concludes that ARCH and GARCH models are effective tools for understanding and forecasting Bitcoin’s volatility. The study underscores the importance of acknowledging volatility persistence and clustering effects when analyzing cryptocurrency price behavior. However, it also highlights areas for improvement in econometric modelling by including the exploration of hybrid models and the integration of macroeconomic factors to enhance forecasting accuracy.Yayın Determinants of Bitcoin price movements(Suat Teker, 2024-07-30) Teker, Dilek; Teker, Suat; Demirel, EsinPurpose- Investors want to include Bitcoin in their portfolios due to its high returns. However, high returns also come with high risks. For this reason, the volatility prediction of Bitcoin prices is the focus of attention of investors. Because Bitcoin's volatility is used as an important input in portfolio selection and risk management. This means that the models to be used in predicting Bitcoin volatility increases the importance of performance. In this research; A comparative examination of the models applied for Bitcoin shows an effective performance in volatility prediction. It is very important for evaluation. The aim of this study is to model Bitcoin price returns and to examine future return predictions and return directions using historical Bitcoin prices. Methodology- Many models have been used in studies on financial instruments and price predictions. Models such as linear and nonlinear regression, Random Walk Model, GARCH and ARIMA fall into this category. Nonlinear econometric models such as ARCH and GARCH are used for financial time series with variable volatility. These models assume that the variance is not constant. In this study, first Bitcoin price returns for the period between January 2020 and December 2023 will be modeled with the GARCH model, and then the ARCH-GARCH models will be used for future prediction of returns for the period between January 2024 and June 2024. Finally, the actual values will be compared with the forecasted values. In other words, the primary aim of this study is to use the daily Bitcoin closing price between May 2020 and December 2023 to estimate the returns for the periods of 2024 and compare it with the actual returns. Findings- The analysis reveals that GARCH Model results showed that in the mean and variance equations, it is seen that all variables are except intercept of the mean equation significant according to the error level of 0.05. Namely, the reaction and persistence parameters are significant accourding to 0.05 in the variance equation. Both the coefficient of the reaction parameter and the coefficient of the persistent parameter are higher than zero (positive). Also, the coefficient of the reaction parameter plus the coefficient of the persistent parameter approximately equals 0.72. That is, it is lower than 1 and higher than zero (positive). The level of persistence is not too high. So, we do not think about non-stationary variance in the model. Reaction parameter’s coefficient is 0.13. And persistence parameter’s coefficient is 0.58. As we can see, persistent parameter is much higher than reaction parameter. That is, when there is a new shock that creates the persistent parameter, that shock will be in effect for a long time, it will not disappear immediately. That is, a significant part of the shock that occurs in one period flows into the next period. After determining the appropriate mean and variance models, a forecast is made using Automatic ARIMA forecasting for BITCOIN return forecasting. This forecast is made for the first five months of 2024, without adding the actual values of the first five months of 2024 to the data. The program ranks the most appropriate model. The program chose GARCH(3,3) as the most appropriate model in "bitcoin return prediction". Conclusion- The results of the test applied in the study can be summarized that the unit root test results showed that it was necessary to work with return series. GARCH(1,1) model results show when there is a new shock that creates the persistent parameter, that shock will be in effect for a long time, it will not disappear immediately. That is, a significant part of the shock that occurs in one period flows into the next period. According to GARCH automatic forecasting results, the best GARCH model that models Bitcoin return is the GARCH(3,3) model. According to these model results, although the slopes of the actual and forecasted return series move in the same direction, the model remains weak for forecasting. In future studies, it may be recommended to estimate Bitcoin returns with non-linear models.
