Razvoj, ovrednotenje in primerjava odločitvenih modelov za napovedovanje delovanja čistilne naprave

diplomska naloga

Marko Štemberger (Author), Biljana Mileva-Boshkoska (Mentor)

Abstract

V diplomski nalogi raziskujemo, kako lahko odločitvene modele v programu Orange uporabimo za čiščenje odpadne vode. Glavno vprašanje, ki nas zanima, je, kako dobro lahko ti modeli napovedujejo učinkovitost čistilnih naprav za odpadne vode, še posebej ko gre za razmerje med različnimi vhodnimi in izhodnimi parametri. Da bi to ugotovili, smo se lotili kombinacije empirične analize in strojnega učenja. Uporabili smo tri različne algoritme: Neural Networks (NN), Random Forest (RF) in Naivni Bayes (NB). Da bi še dodatno izboljšali naše modele, smo vključili tudi algoritem ReliefF, ki nam je pomagal izbrati tiste spremenljivke, ki najbolj vplivajo na naše rezultate. Glavni cilj naše raziskave je bil razjasniti, kako lahko odločitveni modeli pomagajo pri čiščenju odpadne vode. Končni cilji so bili jasni: ustvariti robustne odločitvene modele, preveriti, kako dobro delujejo, in ugotoviti, katere spremenljivke so ključne za uspešno čiščenje odpadne vode.

Keywords

strojno učenje;nevronske mreže;naključni gozd;naivni Bayes;algoritem ReliefF;odločitveni modeli;Orange;

Data

Language:	Slovenian
Year of publishing:	2023
Typology:	2.11 - Undergraduate Thesis
Organization:	FIŠ - Faculty of Information Studies
Publisher:	[M. Štemberger]
UDC:	004.8(043.2)
COBISS:	174939907
Views:	435
Downloads:	32
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary abstract:	In this thesis, we explore how decision models in Orange can be used for wastewater treatment. The main question we are interested in is how well these models can predict the performance of wastewater treatment plants, especially when it comes to the relationship between different input and output parameters. In order to find out, we have undertaken a combination of empirical analysis and machine learning. We used three different algorithms: Neural Networks (NN), Random Forest (RF) and Naive Bayes (NB). In order to further improve our models, we also included the ReliefF algorithm, which helped us to select those variables that have the most impact on our results.The main objective of our research was to clarify how decision models can help in wastewater treatment. The final objectives were to create robust decision models, to test how well they work, and to find out which variables are key to successful wastewater treatment.
Secondary keywords:	machine learning;neural networks;random forest;naive Bayes;ReliefF algorithm;decision models;Orange;Strojno učenje;Nevronske mreže (računalništvo);
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Fakulteta za informacijske študije v Novem mestu
Source comment:	Na ov.: Diplomska naloga : visokošolskega strokovnega študijskega programa prve stopnje;
Pages:	XV, 83 str.
ID:	21492972