Ivana Nizetic Kosovic

Prof. Ivana Nizetic Kosovic

reseracher

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
31 August 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Jelena NakićJelena Nakić
Jelena Nakić
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
AFAna Franić
Ana Franić
Visualization of geographic data is part of many widely used solutions that aim to communicate the information to the end user. Effective visualizations are those that are tailored to a specific group of users and their tasks, as well as to specific properties of the data. Usability is a key concept for such solutions, and the most effective way to achieve high usability is to incorporate user-centered design (UCD) into the development process. In visualization, data are often coded by colors, so the choice of color schemes and color range is critical to enable the reverse mapping of colors to data. In this paper, we present a method for integrating the principles of UCD into the development process. In doing so, we use a case involving the selection of the most appropriate color scheme and associated parameters for representing temperature values on geographic maps. The method described is suitable for use in UCD-based research related to the visualization of various types of data and is useful for researchers and developers of geovisualizations.

26 October 2021 in Water
Verified
Journal Article
Open access
PDF file
DDDaniela Džal
Daniela Džal
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
TMToni Mastelic
Toni Mastelic
DIDamir Ivanković
Damir Ivanković
TPTatjana Puljak
Tatjana Puljak
SJSlaven Jozić
Slaven Jozić
Predictive models of bathing water quality are a useful support to traditional monitoring and provide timely and adequate information for the protection of public health. When developing models, it is critical to select an appropriate model type and appropriate metrics to reduce errors so that the predicted outcome is reliable. It is usually necessary to conduct intensive sampling to collect a sufficient amount of data. This paper presents the process of developing a predictive model in Kaštela Bay (Adriatic Sea) using only data from regular (official) bathing water quality monitoring collected during five bathing seasons. The predictive modelling process, which included data preprocessing, model training, and model tuning, showed no silver bullet model and selected two model types that met the specified requirements: a neural network (ANN) for Escherichia coli and a random forest (RF) for intestinal enterococci. The different model types are probably the result of the different persistence of two indicator bacteria to the effects of marine environmental factors and consequently the different die-off rates. By combining these two models, the bathing water samples were classified with acceptable performances, an informedness of 71.7%, an F-score of 47.1%, and an overall accuracy of 80.6%.

7 March 2023 in Molecules
Verified
Journal Article
Open access
PDF file
Maja Veršić BratinčevićMaja Veršić Bratinčević
Maja Veršić Bratinčević
ABAna Bego
Ana Bego
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
Maja Jukić ŠpikaMaja Jukić Špika
Maja Jukić Špika
FBFilipa Burul
Filipa Burul
EVElda Vitanovic
Elda Vitanovic
The objective of this study was to evaluate the release rate, duration, and biological efficiency of yeast volatile compounds associated with olive fruit flies in slow-release dispensers, polypropylene vials, and rubber septa attached to yellow sticky traps under different environmental conditions in order to protect the environment, humans, and nontarget organisms. Isoamyl alcohol, 2-octanone, and 2-phenethyl acetate were placed in dispensers and tested over a four-week experiment. The weight loss of the volatile compounds in both dispensers was measured, and a rapid, inexpensive, and simple HS-GC/FID method was developed to determine the residual amount of volatiles in the septa. 2-Phenethyl acetate stood out in the rubber septa and showed a statistically significant difference in the release ratio compared to the other volatiles under all conditions tested. Our results showed that the attraction of olive fruit flies increased with decreasing concentrations of the tested volatiles. Regarding the number of flies attracted by rubber septa containing 2-phenethyl acetate, significantly better results were obtained than for septa containing isoamyl alcohol and 2-octanone, in contrast to the attraction of olive fruit flies to polypropylene vials containing these compounds but without significant difference. Since the presence of all tested chemicals was detected during the experiment, this opens the possibility of using more environmentally friendly and cost-effective dispensers with a significantly lower amount of semiochemicals.

3 July 2023 in International Journal of Data Science and Analytics
Verified
Journal Article
Open access
PDF file
JGJelena Čulić Gambiroža
Jelena Čulić Gambiroža
TMToni Mastelic
Toni Mastelic
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
MCMario Cagalj
Mario Cagalj
With the increase in number and size of Internet of Things systems, there is an ever-growing risk of (meta)data loss, as well as the maintenance overhead to mitigate such risks. The experts recognize three main challenges in this area that need to be tackled, namely (1) downsizing the manual work required for configuring sensor networks, (2) recovering metadata, such as sensor type, in case of connection issues, malfunctions or malicious actions in sensor networks, (3) rebuilding metadata lost due to unexpected problems within a data storage. Fortunately, all three challenges can be tackled with a uniform solution, namely the signal type classification approach, which is able to match raw signal to an appropriate data type. In this research, we evaluate and compare different approaches for signal type classification that can be used to recognize a signal type being read from an IoT sensor. This is done by using machine learning methods for modelling a signal represented as raw time series data. Three machine learning classification approaches are taken into a consideration, namely one class, two class and multi-class. According to the results of the evaluation, the most accurate multi-class random forest algorithm can correctly classify unknown signals in $$\sim {75}\%$$ ∼ 75 % of the cases based on only 20 consecutive sensor readings. Moreover, multi-class random forest can detect two most probable classes of monitored signal with the accuracy of 95%.

27 March 2022
Verified
Conference Paper
Open access
DKDiana Škurić Kuraži
Diana Škurić Kuraži
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
IBIvana Herceg Bulić
Ivana Herceg Bulić
Forest fire research can comprise forest fire case studies, laboratory experiments, fire detection by ground sensors, unmanned aerial vehicles and satellites, development of fire behaviour models, fire danger forecast, fire risk assessment, and much more. Commonly used and accepted Canadian method for forest fire danger forecast is expressed as Fire Weather Index (FWI) uses weather data. The index estimates the danger of wildfire and is based on meteorological parameters (air temperature, air humidity, wind speed, and rainfall amount) referring to 12 UTC for that day at the meteorological station or on a numerical weather prediction model grid point.Knowing how weather and soil interact and affect each other, we propose a new fire risk index based on the innovative Soil Index. Using open-access data, we collected different soil data such as soil temperature and soil moisture, land cover, vegetation, slope, etc. Since there are different types of vegetation and states, Leaf Area Index (LAI) and Normalized Difference Vegetation Index (NDVI) are considered as well. Being focused on forest fires, data about the burned area were also taken into account as well as the slope of the terrain for which the fire risk is calculated.Since all mentioned data have a diverse horizontal and temporal resolution, we decided to group them by temporal resolution: static, semi-static, and dynamic data. Static data refers to data that rarely change (never or every few years; e.g. land cover). Semi-static data refers to data that vary weekly or monthly (e.g. LAI). Dynamic data group refers to data that is strongly influenced by weather conditions (like soil temperature) and varies every hour. Because of various horizontal resolutions, soil parameters are interpolated to the same horizontal grid. Soil parameters are analysed concerning historical forest fires in Croatia. Despite Soil Index being based on soil parameters, we compared it with Fire Weather Index using data records for historical forest fires in Croatia. Obtained results indicate that the soil index has a better prediction performance compared to FWI. This study also highlights that not only the meteorological environment but also soil conditions are important parameters for fire risk assessment.

1 September 2022 in Journal of Communications Software and Systems
Verified
Journal Article
Open access
PDF file
IKIvana Nizetic Kosovic
Ivana Nizetic Kosovic
TMToni Mastelić
Toni Mastelić
DSDomina Sokol
Domina Sokol
DKDiana Škurić Kuražić
Diana Škurić Kuražić
Spatio-temporal modelling is an emerging research area due to the increasing availability of sensor data collected across space and time. The models are build either with a model-driven or data-driven approach. The former often results in complex monolith models that are not suitable for lightweight Edge deployment. The latter requires a vast amount of data and may not provide an overall good performance. Consequently, the data-driven approach is being used to substitute only parts of model-driven outputs, by creating micromodels that tackle spe-cific scenarios. The main contribution of this paper is a definition and demonstration of the process for finding such scenarios for which a spatio-temporal model could be improved or replaced by a micromodel and deployed on Edge. The process is demonstrated on an example of a Numerical Weather Prediction model (NWP), namely its outputs of temperature and precipitation. NWP is evaluated using black-box testing considering the specificity of spatial and temporal components, in both normal and extreme conditions. The novelty of this process is its ability to highlight weaknesses of the existing expert models and suggest scenarios in which the models can be improved and deployed on the Edge.