Endüstri Mühendisliği Bölümü Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.11779/1942
Browse
2 results
Search Results
Article Citation - WoS: 5Citation - Scopus: 12Predicting Cash Holdings Using Supervised Machine Learning Algorithms(Springer, 2022-05-18) Özlem, Şirin; Tan, Ömer FarukThis study predicts the cash holdings policy of Turkish firms, given the 20 selected features with machine learning algorithm methods. 211 listed firms in the Borsa Istanbul are analyzed over the period between 2006 and 2019. Multiple linear regression (MLR), k-nearest neighbors (KNN), support vector regression (SVR), decision trees (DT), extreme gradient boosting algorithm (XGBoost) and multi-layer neural networks (MLNN) are used for prediction. Results reveal that MLR, KNN, and SVR provide high root mean square error (RMSE) and low R2 values. Meanwhile, more complex algorithms, such as DT and especially XGBoost, derive higher accuracy with a 0.73 R2 value. Therefore, using advanced machine learning algorithms, we may predict cash holdings considerably.Article Citation - WoS: 9Citation - Scopus: 10Sequential Testing in Batches(Springer, 2016-10-04) Ünlüyurt, Tonguc; Shahmoradi, Zahed; Özluk, Özgur; Selcuk, Barış; Daldal, RebiWe study a new extension of the Sequential Testing problem with a modified cost structure that allows performing of some tests in batches. As in the Sequential Testing problem, we assume a certain dependence between the test results and the conclusion. Namely, we stop testing once a positive result is obtained or all tests are negative. Our extension, motivated by health care applications, considers a fixed cost associated with executing a batch of tests, with the general notion that the more tests are performed in batches, the smaller the total contribution of fixed costs to the sequential testing process. The goal is to minimize the expected cost of testing by finding the optimal choice and sequence of the batches available. The resulting NP-hard model is a variation of the set partitioning problem. We propose various heuristic algorithms for the effective solution of the problem and then demonstrate the performances of the algorithms through extensive numerical experiments.