Forecasting of production program

V magistrski nalogi je predstavljena problematika napovedovanja. V teoretičnem delu je najprej prikazan proces in pomen napovedovanja, ter dejavniki, ki vplivajo na napovedovanje. Sledi podroben pregled kvantitativnih ter kvalitativnih metod napovedovanja. V nadaljevanju teoretičnega dela so predstavljeni načini merjenja natančnosti napovedovanja. Teoretičnemu delu magistrske naloge sledi praktičen prikaz izdelave napovedi. Glavni cilj magistrske naloge je bil izdelati napovedi z uporabo izbranih kvalitativnih metod napovedovanja na podlagi realnih podatkov v obliki časovnih vrst. Metode, ki smo jih uporabili pri napovedovanju so: napovedi iz linearnega trenda, metoda drsečega povprečja, metoda uteženega drsečega povprečja in metoda enostavnega eksponentnega glajenja. Z uporabo teh štirih izbranih metod smo dobili dvanajst različnih napovedi za tri izbrane primere. Rezultate napovedovanja smo primerjali z dejanskimi podatki in tako določili odstopanja in posledično natančnost posameznih metod napovedovanja. Na podlagi medsebojne primerjave odstopanj in napak pri posameznih metodah smo lahko določili najprimernejšo metodo za napovedovanje. Le ta je v našem primeru metoda uteženega drsečega povprečja. Ta metoda se je namreč izkazala kot najbolj natančna pri vseh treh primerih napovedovanja na podlagi časovnih vrst realnih podatkov. Ob koncu smo za vsak realni primer napovedovanja razvrstili metode po vrstnem redu glede na natančnost.The Master\u27s thesis examines the problems of forecasting. In the beginning of the theoretical part, the forecasting, its importance, the process and factors that affect the forecasting are presented. This section is followed by detailed examination of quantitative and qualitative methods of forecasting. Furthermore, some ways of measuring the accuracy of forecasting are described. The theoretical part of Master’s thesis is followed by a practical demonstration of making forecasts. The main objective of the Master’s thesis was to make a forecast using selected qualitative forecasting methods that are based on the real data in the form of time series. The methods we used to make forecasts are: linear trend line method, moving average method, weighted moving average method and exponential smoothing method. By using these four selected methods we got twelve different forecasts for three selected cases. We compared the results of forecasts with actual data, defined deviation and, consequently, the accuracy of individual forecasting method. The most appropriate method for the forecast was selected based on the comparison of deviations and errors in different methods. The most appropriate method in this case is the method of weighted moving average. It was proved to be the most accurate method in all three cases of forecasting based on time series of real data. In the end, we ranked all methods in order of the accuracy for each real case of forecasting