Metin, S.K.Taze M.Uymaz, H.A.Okur E.2023-06-162023-06-1620179.78E+12https://doi.org/10.1109/SIU.2017.7960327https://hdl.handle.net/20.500.14365/360225th Signal Processing and Communications Applications Conference, SIU 2017 -- 15 May 2017 through 18 May 2017 -- 128703Detection of multiword expressions is an important pre-task in several research topics such as natural language understanding, automatic text summarization, and machine translation in the area of natural language processing. In this study, detection of multiword expressions in Turkish texts is accepted as a classification problem. 6 types of linguistic features are defined solving this problem in Turkish texts. The classification tests are performed by 10 different classifiers utilizing the prepared data set. The performance of classifiers is measured for different sizes of random train-test sets by running the tests 10 times. The test results showed that linguistic features can be used in identification of multiword expressions. And it is observed that SMO and J48 algorithms reached the highest classification performances based on different evaluation metrics. © 2017 IEEE.trinfo:eu-repo/semantics/closedAccessclassification problemlinguistic featuresmultiword expression detectionnatural language processingFeature extractionLinguisticsNatural language processing systemsProblem solvingSignal processingStatistical testsAutomatic text summarizationClassification performanceClassification testsLinguistic featuresMachine translationsMulti-word expressionsNatural language understandingPerformance of classifierClassification (of information)Conference Object10.1109/SIU.2017.79603272-s2.0-85026309175