Análisis inteligente del mercado laboral en Ecuador: modelo de machine learning con datos abiertos de economía y empleo.

Abstract

Abstract: This paper develops a predictive model based on Machine Learning to analyze the Ecuadorian labor market by integrating open data from the National Institute of Statistics and Censuses (INEC) and the Central Bank of Ecuador. Using the CRISP-DM methodology, 21,049 records were consolidated, relating macroeconomic variables (country risk, oil price, public debt) with labor indicators, divided by demographic and sectoral characteristics. A feature engineering method was applied to maintain a history of the macroeconomic variables, recording the state of each variable for 1, 3, 6, and 12 months. Six models were evaluated: Linear Regression, Ridge, Random Forest, XGBoost, Gradient Boosting, and SVR, to compare the results and select the one with the best performance for each test. The importance analysis of variables validated hypotheses of the National Development Plan and demographic characteristics to identify the most relevant variables. The model provides the ability to anticipate trends up to 3 months in advance. An executive dashboard was implemented that democratizes access to results for non-technical decision-makers.