20.500.12556/RUL-163764 Dinamično gručenje modelov vozil glede na ceno in dohodek gospodinjstev: Čezdržavna analiza za napredne modele napovedovanja trga magistrsko delo Dynamic clustering of vehicle models based on price and household income: A cross-country analysis for enhanced market prediction models Obstoječa klasifikacija proizvajalce razvršča v eno od štirih kategorij: ekonomska, količinska, premijska ali luksuzna vozila. Trg motornih vozil postaja vse bolj kompleksen in proizvajalci vozil predstavljajo nove modele zunaj svojih tradicionalnih niš. Zato je vse težje enemu proizvajalcu dodeliti eno samo kategorijo. Namen te naloge je bil razviti nov pristop k razvrščanju, ki namesto proizvajalcev razvršča vozila. Razvita sta bila dva modela: eden razvršča vozila znotraj posameznih držav, drugi pa znotraj držav in segmentov. Za čim boljšo predstavitev stanja na trgu, položaja proizvajalcev in značilnosti vozil so bili parametri za razvrščanje skrbno izbrani. Ti vključujejo ceno, dohodek gospodinjstva, BDP, ugled proizvajalca, izdatke za avtomobile, obliko vozil in segment. Oba modela razvrščanja vozila razdelita v šest kategorij. Z modeliranjem različnih scenarijev modeli omogočajo globlji vpogled v trende na trgu. Uporaba algoritmov za razvrščanje pomaga pri načrtovanju prodajnih strategij in omogoča uporabo rezultatov v drugih modelih napovedovanja. The existing classification system assigns manufacturers to one of four categories: Economy, Volume, Premium, or Luxury. However, as the market evolves and brands offer vehicles outside their traditional niches, it becomes increasingly difficult to assign a single category to a manufacturer. The goal of this thesis was to develop a new classification approach that addresses this by clustering vehicles instead of manufacturers. Two models were developed: one clusters vehicles within countries, and the other clusters vehicles within both countries and segments. To summarize the state of the market, the position of manufacturers, and vehicle characteristics, the cluster parameters were carefully selected. These include price, household income, GDP, manufacturer reputation, automotive expenditures, vehicles bodystyle, and segment. Both clustering models classify cars at the chosen level into six categories. By simulating different scenarios, the models provide deeper insights into market trends. The use of clustering algorithms helps in strategies decision-making processes and allows the results to be applied in other forecasting models. klasifikacija strojno učenje tržni trendi v avtomobilski industriji analiza cen napovedno modeliranje analiza podatkov vpogledi v trg napovedni modeli clustering machine learning automotive market trends price analysis predictive modeling data analysis market insights forecasting models true false false Angleški jezik Slovenski jezik Magistrsko delo/naloga 2024-10-10 11:45:56 2024-10-10 11:46:01 2024-10-12 03:51:53 0000-00-00 00:00:00 2024 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 2025-10-07 519.8 144795 210512387 15002.pdf 15002.pdf 1 010E78E3BC6896632F6567286ED63A30 c24698460117da56ab58595f450084f3f94b1dd5afa2e5d46864cd8bcf9dc851 46ab8c85-86ec-11ef-b232-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=193671 Fakulteta za matematiko in fiziko 0 0 0