Miloš Todorov

Dr. Miloš Todorov

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 June 2024 in Journal of Hydrologic Engineering
Verified
Journal Article
Closed access
IKIvana Gojko Krtolica
Ivana Gojko Krtolica
MTMiloš Todorov
Miloš Todorov
SPStevan Prohaska
Stevan Prohaska
MSMilan Stojkovic
Milan Stojkovic

25 April 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
Ninoslava TihiNinoslava Tihi
Ninoslava Tihi
MTMiloš Todorov
Miloš Todorov
SPStefan Mirko Popovic
Stefan Mirko Popovic
Srdjan PopovSrdjan Popov
Srdjan Popov
NSNiko Samec
Niko Samec
Filip KokaljFilip Kokalj
Filip Kokalj
Natural disasters, including droughts, exert a profound impact on public health, agriculture, industry, and the environment. The intensification of climate change has exacerbated the frequency and severity of such events. In the context of smart cities, the implementation of advanced prediction methodologies is pivotal for mitigating or even preventing these catastrophic outcomes. Among the most widely employed techniques for time series forecasting are the (S)ARIMA and Holt-Winters exponential smoothing methods, which form the foundation of this study. This research proposes a robust framework for the preprocessing of raw meteorological data, the application of forecasting techniques to monthly observations, and the evaluation of prediction models using standardized performance metrics. The objective is to identify the optimal model for predicting key parameters - such as air temperature, precipitation, and soil moisture - that serve as critical indicators of urban drought conditions. The findings reveal that the SARIMA model demonstrates superior reliability for time series exhibiting a decreasing trend, such as soil moisture, while the Holt-Winters method excels in modeling time series with no trend (e.g., precipitation) and those characterized by strong seasonality, such as air temperature. These insights contribute to the development of more effective urban drought prediction systems and highlight the nuanced applicability of these forecasting methods in addressing climate-related challenges.

8 April 2026 in Algorithms
Verified
Journal Article
Open access
PDF file
MTMiloš Todorov
Miloš Todorov
Darjan KarabasevicDarjan Karabasevic
Darjan Karabasevic
Predrag M. TekicPredrag M. Tekic
Predrag M. Tekic
DDDragana Dudić
Dragana Dudić
DVDejan Viduka
Dejan Viduka
This paper introduces a new hybrid forecasting architecture that combines Nonlinear Autoregressive (NAR) models, the proxy Standardized Precipitation-Evapotranspiration Index (SPEI), and a Nonlinear Autoregressive with Exogenous Inputs (NARX) framework for Soil Moisture Index (SMI) prediction. The suggested methodology solves the crucial difficulty of combining future climatic knowledge into soil moisture forecasting by using a cascaded approach. Stage 1 uses univariate NAR models to create multi-step-ahead predictions of precipitation and temperature. Stage 2 converts these forecasts into proxy SPEI values using a physically based water balance computation, and Stage 3 employs a NARX model that uses observed historical SMI and forecast-derived proxy SPEI as exogenous inputs. The framework is assessed using high-frequency observations from 2014 to 2020, with training data through 2019 and validation covering the whole 2020 horizon. The study combining forecast-driven climatic indicators with autoregressive soil moisture dynamics resulted in prediction accuracy (R2 = 0.9888, RMSE = 0.0827, MAE = 0.0567). This study presents a new NAR–SPEI–NARX model for SMI prediction forecasting, based on three-stage modeling, where NAR models forecast precipitation and temperature and then turn them into SPEI-proxy as an exogenous input for NARX.

1 January 2024
Verified
Conference Paper
Closed access
MCMiroslava Mihajlov Carević
Miroslava Mihajlov Carević
MTMiloš Todorov
Miloš Todorov
Real-world problems necessitate intricate solutions that encompass the combination of science, technology, engineering, arts, and mathematics. Consequently, change has occurred in educational paradigms, resulting in the adoption of the STEAM method in education. The STEAM approach combines the disciplines of science, technology, engineering, art/architecture, and mathematics, integrating them into a cohesive framework that offers a fresh method for addressing challenges within the educational system. The implementation of STEAM practices has garnered significant attention, particularly in the realm of higher education. They include equipping students with interdisciplinary skills necessary for the job market. Additionally, they contribute to the development of students intellectual curiosity. Across the world, STEAM education has become a widely used word. Nevertheless, it is imperative to do research and enhance the implementation of STEAM in both classroom settings and teacher education. This paper showcases instances of interdisciplinary instruction implemented at Alfa BK University in Belgrade. The initial example illustrates the process of directing students towards the integration of mathematics, engineering, and the arts to attain a realization of harmony and beauty in architecture and art. The second example demonstrates the incorporation of the data from the Statistical Office of the Republic of Serbia using mathematical techniques, computer programming, and visual representation.

1 January 2025
Verified
Conference Paper
Closed access
Ninoslava TihiNinoslava Tihi
Ninoslava Tihi
MTMiloš Todorov
Miloš Todorov
MPMarko Pavlović
Marko Pavlović
Filip KokaljFilip Kokalj
Filip Kokalj
In the age of smart cities, where interconnected technologies produce substantial data, time series forecasting has become a crucial instrument for efficient urban management. Exploratory Data Analysis (EDA) is an essential initial phase in comprehending and organising data for precise and insightful predictions. This study examines the utilisation of exploratory data analysis approaches for time series data in the framework of smart cities. It emphasises techniques for detecting trends, seasonality, autocorrelations, and correlations in data streams from sources like environmental monitoring systems. The research investigates ways to identify patterns, test hypotheses, and confirm assumptions using visual and quantitative methods. Employing diverse visualisation tools, statistical summaries, and decomposition techniques enhances the comprehension and preprocessing of time series datasets. The objective of the research is to perform and evaluate the efficacy of standard tools such as time plots, autocorrelation and correlation analysis, seasonality decomposition, and identifying trend patterns in conjunction with sophisticated techniques such as seasonal decomposition of time series (STL), correlation heatmaps, and trend detection techniques. The findings emphasise EDA’s significance as a fundamental step in empowering researchers and practitioners to make informed decisions, ensuring strong analytical results, and developing resilient time-series forecasting models for contemporary urban issues.