Napovedovanje količine odstrela divjih živali na Primorskem

diplomsko delo

Iza Jerman Slabe (Author), Lovro Šubelj (Mentor)

Abstract

Lovska Zveza Slovenije in Zavod za gozdove vsako leto določita število predvidenga odstrela divjadi v Sloveniji. Da bi lovci, katerih število z leti upada, lažje dosegali rezultate, je bil razvit program za odločanje, ali naj se ob določenih pogojih odpravijo na lov. Podatkom o dnevnem številu odstrela v posameznih loviščih so bili pridruženi podatki o vremenu in dodatne informacije o lovskih družinah, razmerah ter divjadi. Iz velikega nabora atributov je izbrana podmnožica najpomembnejših in na tako izbranih podatkih s pomočjo pogosto uporabljenimi metodami strojnega učenja narejen model odločanja. Uporabljene metode so metoda $k$-najbližjih sosedov, linerna regresija, metoda podpornih vektorjev, umetne nevronske mreže in naključni gozdovi. Za najbolj uspešen model se je izkazalo napovedovanje z naključnimi gozdovi, ki predvidi, ali se je primerno za vikend odpraviti na lov. Za atribute, ki najbolj vplivajo na odločitev odprave na lov, sta se izkazala veter in pojav megle.

Keywords

regresija;lovstvo;računalništvo in matematika;interdisciplinarni študij;univerzitetni študij;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[I. Jerman Slabe]
UDC:	004.85:639.1(043.2)
COBISS:	78960387
Views:	735
Downloads:	61
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Predicting hunting success of wild animals in Primorska region
Secondary abstract:	Every year, the Hunters’ Association of Slovenia and the Slovenia Forest Service determine the annual cull. As the number of hunters is declining, a program has been developed to help them achieve the expected results. The program helps with their decision whether to go hunting on a specific day based on existing and predicted conditions. Data about the daily number of culls in individual hunting areas was combined with the data about time of the culls and additional information about hunting families, conditions and game. From a large set of attributes, a subset with the most important attributes is selected. On that subset, a decision-making model is made with the help of the most frequently used machine learning methods. Used methods are $k$-nearest neighbours, linear regression, support vector regression, artificial neural networks, and random forests. The most successful model was predicted with random forests which informs whether hunters should go hunting for the weekend. Attributes which best define the decision to go hunting proved to be wind strength and the presence of fog.
Secondary keywords:	machine learning;regression;hunting;computer science;computer and information science;computer science and mathematics;interdisciplinary studies;diploma;Strojno učenje;Lov;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS):	Bachelor thesis/paper
Study programme:	1000407
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	44 str.
ID:	13540291