Dissertation Title: ASSESSMENT OF NON-TECHNICAL SKILLS THROUGH THE APPLICATION OF MACHINE LEARNING ALGORITHMS RECOMMENDATIONS FOR PUBLIC EMPLOYMENT INSTITUTIONS IN ALBANIA

Authors: Milena sheikh

• Institution: University of Tirana,  Faculty of Economics, Department of Statistics and Applied Informatics

• Field of study:  Applied Statistics and Informatics / Applied Statistics and Mathematics in Economics

• Publication date: 15.06.2026

• The dissertation is published in Albanian.

© Copyright: Milena sheikh

Published by the University of Tirana
Based on legal acts, regulations and policies of the UT

👉 Click here to view the full dissertation (PDF)

 

Abstract

Technological transformations and globalization have significantly increased the complexity of the labor market, making the identification and analysis of the skills required by job seekers a methodological and institutional challenge.

This dissertation focuses on the use of Topic Modeling (TM) algorithms for the automatic analysis of job vacancy announcements in Albania, with particular emphasis on the application of the Latent Dirichlet Allocation (LDA) algorithm for identifying thematic structures and non-technical skill patterns.

First, the main thematic modeling algorithms, including LSA, NMF, LDA and BERTopic, are theoretically analyzed, analyzing their advantages and limitations in the conditions of Albanian text data. Next, LDA is analyzed in detail from a statistical perspective and applied on a corpus of job advertisements to extract coherent topics and dominant skills required by the Albanian labor market,

The main practical contribution of the dissertation is the development of the “Skills Map” application, built in the R Shiny language. The application enables the loading and pre-processing of job descriptions, the analysis of only the requirements and job description sections, the identification and categorization of non-technical skills, by applying the LDA algorithm and linguistic processing with UDPipe for the Albanian language.

Field: Statistics and Applied Mathematics in Economics

Keywords: Latent Dirichlet Allocation, Topic Modeling, Non-technical skills, Natural Language Processing, Labor market in Albania

 

Abstract

Technological transformations and globalization have significantly increased the complexity of the labor market, turning the identification and analysis of necessary skills into both a methodological and institutional challenge.

This study examines the use of topic modeling algorithms for the automated analysis of job vacancy postings in Albania, with a particular focus on applying the Latent Dirichlet Allocation (LDA) algorithm to uncover thematic structures and patterns related to non-technical skills. The study first provides a theoretical analysis of the main Topic Modeling algorithms, including LSA, NMF, LDA, and BERTopic, examining their advantages and limitations when applied to a corpus with text data written in the Albanian language. Further in the thesis, LDA is analyzed in detail from a statistical perspective and applied to a corpus of job advertisements to extract coherent topics and dominant skills demanded by the Albanian labor market.

The main practical contribution of the dissertation is the development of the "Skill Map" application, built using R Shiny. The application enables users to upload and preprocess job vacancy descriptions, analyze only the requirements and job description sections, identify and categorize non-technical skills, by applying the LDA algorithm and UDPipe for the Albanian language.

Field of study: Statistics and Applied Mathematics in Economics

Keywords: Latent Dirichlet Allocation, Topic Modeling, Soft Skills, Natural Language Processing, Albanian Labor Market