magistrsko delo
Tatyana Unuchak (Author), Mirjana Kljajić Borštnar (Mentor)

Abstract

Namen te raziskave je ugotoviti, kako lahko uporaba tehnik strojnega učenja izboljša procese upravljanja nalog v projektnih dejavnostih podjetij. Študija je bila izvedena z uporabo metodologije CRISP-DM na podatkih iz dveh virov o projektih Jira. Glavni koraki priprave podatkov so vključevali čiščenje podatkov, tokenizacijo, lematizacijo in uravnoteženje. Modeliranje je bilo izvedeno z uporabo petih klasifikatorjev: Random Forest, SVC (angl. support vector classifier), Logistic Regression, Gradient Boosting in kNN. Upoštevani so bili različni pristopi k razvrstitvi podatkov: razvrstitev v dva, tri in štiri razrede. Analiza je pokazala, da čiščenje podatkov iz tehničnih informacij ne vpliva na rezultate razvrščanja. Uravnoteženje je izboljšalo rezultate. Po našem mnenju je razvrstitev podatkov v dva, tri in celo štiri razrede pokazala dobre rezultate. Uvedba sentimentalne sestavine v model ni izboljšala rezultatov razvrščanja. Menimo, da je bil cilj raziskave dosežen. Nadaljnje raziskave so lahko usmerjene v izboljšanje algoritmov za čiščenje opisov projektnih nalog iz tehničnih informacij. Naši rezultati in priporočila bodo pripomogli k izboljšanju procesov upravljanja nalog v projektih in povečanju njihove učinkovitosti.

Keywords

vodenje projektov;strojno učenje;prioriteta nalog;algoritmi za razvrščanje;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UM FOV - Faculty of Organizational Sciences
Publisher: [T. Unuchak]
UDC: 004.8
COBISS: 200810499 Link will open in a new window
Views: 10
Downloads: 0
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Applying machine learning methods to improve the task management process in project activities
Secondary abstract: The purpose of this research is to identify how the use of machine learning techniques can improve task management processes in companies' project activities. The study was carried out using the CRISP-DM methodology on data from two sources on Jira projects. The main data preparation steps included data cleaning, tokenization, lemmatization and balancing. The modelling was performed using five classifiers: Random Forest, SVC (support vector classifier), Logistic Regression, Gradient Boosting and kNN. Different approaches to data classification were considered: two, three and four class classification. The analysis showed that data cleaning from technical information does not affect the classification results. The rebalancing improved the results. In our opinion, data classification into two, three and even four classes showed good results. Introducing a sentiment component into the model did not improve the classification results. We believe that the purpose of the study has been achieved. Further research can be directed towards improving algorithms for cleaning project task descriptions from technical information. Our results and recommendations will help to improve the task management processes in projects and increase their efficiency.
Secondary keywords: Strojno učenje;Univerzitetna in visokošolska dela;
Type (COBISS): Master's thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za organizacijske vede
Pages: V, 71 f.
ID: 24295390