This bachelor thesis deals with the development of predictive models for the prices of LEGO sets on the secondary market. The first category of models attempts to use various attributes of LEGO sets to predict the current price of an unwrapped, no longer officially sold set on the secondary market. Data from 3608 completed transactions for 1074 different set s are used for this prediction. The observed transactions were conducted through the online marketplace eBay.com in the U.S. territory betw... zobrazit celý abstraktThis bachelor thesis deals with the development of predictive models for the prices of LEGO sets on the secondary market. The first category of models attempts to use various attributes of LEGO sets to predict the current price of an unwrapped, no longer officially sold set on the secondary market. Data from 3608 completed transactions for 1074 different set s are used for this prediction. The observed transactions were conducted through the online marketplace eBay.com in the U.S. territory between September and November 2023. The prediction models are built by regression analysis. The explanatory variable in the models is the current price of the kit on the secondary market. The best model was able to explain 80% of the variance of the current price. The second category of models attempts to predict the future price of sealed and no longer officially sold sets based on the previous price. ARIMA model is used for this prediction. The model succeeds in predicting the prices of building sets on the secondary market six months in advance. |