03905ntm a22006017i 4500 000677008 CZ-PrVSE 20200829134027.0 m d cr n|||||||||| 200829s2020 xr fsbm 000 0 eng d NEZPRACOVANÝ IMPORT ABA006 cze ABA006 ABA006 rda Barus, Miroslav ISIS:112140 dis Application of a machine-learning pipeline to a multiclass classification problem eng Application of a machine-learning pipeline to a multiclass classification problem / Miroslav Barus 2020 ?? stran : digital, PDF soubor Vedoucí práce: Jan Zouhar Diplomová práce (Ing.)—Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky, 2020 Obsahuje bibliografii Textový (vysokoškolská kvalifikační práce) Rok obhajoby 2020 A problem of predicting a future event outcome from a fixed set of possible outcomes can be approached as a traditional machine-learning classification task. There are several steps that need to be done in order to build a reliable machine-learning pipeline with a solid performance on unseen future test data. First, the data should be analyzed using appropriate exploratory data analysis techniques. Various data types need to be pre-processed and transformed, utilizing different feature engineering methods before hypotheses created during the analysis can be tested. During the modeling process, some data can be omitted, while some data might need to be transformed to achieve the best predictive power. Different classification algorithms, such as k-NN, logistic regression, random forest, XGBoost, LightGBM, can be applied with respect to their assumptions and limitations. Hyper-parameter tuning process can serve to improve the predictive performance of individual models. Multiple classification models can be combined using ensemble methods to collectively outperform the best individual model. This thesis addresses all described practices and applies them to the problem of predicting the future interest level in rental apartment listings, which is a multiclass classification problem. The source of the data used for modeling is an already finished competition organized by Kaggle in cooperation with Two Sigma Connect, and Renthop. The final model is a stacked ensemble of previously mentioned models using the LightGBM classifier. The final model score would place among the top 25% of the participants’ model scores. Způsob přístupu: Internet ekonometrie a operační výzkum [obor dipl. práce] diplomové práce fd132022 czenas master's theses eczenas exploratory data analysis feature engineering k-NN logistic regression random forest XGBoost LightGBM hyper-parameter tuning ensemble methods Kaggle multiclass classification machine-learning Zouhar, Jan, 1982- vse2009539908 ths Zimmermann, Pavel, 1980- jo20251268570 opn Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky kn20010709399 dgg https://insis.vse.cz/zp/69934/podrobnosti VŠKP v InSIS https://insis.vse.cz/zp/69934 Hlavní práce https://insis.vse.cz/zp/69934/posudek/vedouci Hodnocení vedoucího https://insis.vse.cz/zp/69934/posudek/oponent/68016 Oponentura https://insis.vse.cz/zp/69934/podrobnosti dc:identifier NEPOSILAT VSKP vse69934 200828 69934