Kosta Mitreski

Diatom Classification with Novel Bell Based Classification Algorithm

Diatoms are ideal indicators of certain physical-chemical parameters and in the relevant literature they are classified into one of the water quality classes (WQCs). Using information technologies methods, we can classify old and new diatoms directly from measured data. In this direction, a novel method for diatom classification is proposed in this paper. The classification models are induced by using modified bell fuzzy membership functions (MFs) in order to make more accurate models. An intensive comparison study of the fuzzy MFs distribution with the proposed method and the classical classification algorithms on the classification accuracy is studied. Based on this evaluation results, three models are presented and discussed. The experimental results have shown that the proposed algorithm remains interpretable, robust on data change and achieve highest classification accuracy. The obtain results from the classification models are verified with existing diatom ecological preference and for some diatoms new knowledge is added.

Predicting chemical parameters of the water from diatom abudance in lake Prespa and its tributaries

In this work, we are modelling the physic-chemical parameters of water using bioindicator data (diatom taxa abundance data). Chemical status of the water (or water quality class) is defined by the values of measured physic-chemical parameters. Traditional approach to model these data is to learn a separate model for each parameter and then derive a global overview with some kind of summarization over the multiple models. Another approach is to learn a single model that describes all parameters (multi target approach). We explore these approaches and apply them on data from Lake Prespa and its tributary rivers. The obtained models revealed interesting connections between the diatom taxa and the water quality (i.e. the values of the chemical parameters).

Influence of Algebraic T-norm on Different Indiscernibility Relationships in Fuzzy-Rough Rule Induction Algorithms

The rule induction algorithms generate rules directly in human-understandable if-then form, and this property is essential of successful intelligent classifier. Similar as crisp algorithms, the fuzzy and rough set methods are used to generate rule based induction algorithms. Recently, a rule induction algorithms based on fuzzy-rough theory were proposed. These algorithms operate on the well-known upper and lower approximation concepts, and they are sensitive to different T-norms, implicators and more over; to different similarity metrics. In this paper, we experimentally evaluate the influence of the T-norm Algebraic norm on the classification and regression tasks performance on three fuzzy-rough rule induction algorithms. The experimental results revealed some interesting results, moreover, the choice of similarity metric in combination with the T-norm on some datasets has no influence at all. Based on the experimental results, further investigation is required to investigate the influence of other T-norms on the algorithm’s performance.

GIS Flood Prediction Models of “Kriva Reka” River

Floods are a natural phenomenon that can cause damage on town building, villages and farmlands, by increasing the water level of nearby river or river systems. The work in this paper aims to present the GIS flood prediction model for the “Kriva Reka” River. By providing early warning about the heavy rain from the national meteorological institute, in combination with our GIS flood prediction model, it will be possible to reduce the damage caused by the floods. The model contains analysis of the terrain data, the hydrometeorological data, and visualizing the geographic river map of flooded areas. We provided the GIS prediction model with the necessary terrain data and hydro- meteorological data for a 5 years period. The visual results from the GIS model show critical areas, where in period of heavy rain, they are potential disaster zones. In the future, we plan to upgrade the GIS system to be available for the citizens via mobile platform, so we can increase the public awareness of such events and help public evacuation.

Diatom Indicating Property Discovery with Rule Induction Algorithm

In the relevant literature the diatoms have ecological preference organized using rule, which takes into account the important influencing physical-chemical parameters on the diatom abundance. Influencing parameters group typically consist from parameters like: conductivity, saturated oxygen, pH, Secchi Disk, Total Phosphorus and etc. In this direction, this paper aims in process of building diatom classification models using two proposed dissimilarity metrics with predictive clustering rules to discover the diatom indicating properties. The proposed metrics play important rule in this direction as it is in every aspects of the estimating quality of the rules, from dispersion to prototype distance and thus lead to increasing the classification descriptive/predictive accuracy. We compare the proposed metrics by classification and rule quality metrics and based on the results, several set of rules for each WQ and TSI category classes are presented, discussed and verified with the known ecological reference found in the diatom literature.

Fuzzy models with GIS for water quality diatom-indicator classification

The level of certain or set of physico-chemical parameter(s) determinates the optimal living condition for the organism. These thresholds in biology are express using categories of classes. One such category which consists from five classes is water quality (WQ) category based on Saturated Oxygen. Due to natural or human-made pressure on the ecosystem, bottom line as that we have to monitor the levels. Many of the diatoms are very sensitive to certain parameters and for these diatoms ecological indicator reference are known in the literature. But, there are many other unknown ecological references to be identified especially for the newly discovered diatoms. The diatom measurement data is noisy, usually comprises from several dimensions and also has non-linear relationship between the environmental parameters and the diatoms abundance. All these properties of the diatom dataset do not meet the assumptions of conventional statistical procedures. To overcome these problems, this research aims to a fuzzy classification approach to describe this diatom-indicator relationship. Using the approach, each diatom will be classified and together with the Geographic Information System (GIS) software, will be represented on map based on Saturated Oxygen parameter. The results from the model are verified with the known ecological reference found in the literature. Even more important, the proposed method has added some new ecological references that do not exist. This fuzzy approach can be modified for defining new WQ category classes based not only for Saturated Oxygen parameters, but also for metals and other physico-chemical parameters.

Novel Inverse Sigmoid Fuzzy Approach for Water Quality Diatom Classification

The prediction accuracy of the fuzzy diatom models depends on both the manner of defining the fuzzy sets used (their number, shape and the parameters of the membership function (MF)) and the kind of the similarity metric used. In this paper, we define new similarity metric, which takes into the account the maximum number of diatoms’ abundance in specific environmental parameter range. The inverse sigmoid MF is used to shape each MF to describe this relationship, in order to produce more accurate models. This improvement of the ecological modelling is achieved through the process of evaluation results for interpretability; higher prediction accuracy and over fitting resistant. The evaluation results compared with classical classification algorithms have confirmed these findings. Based on these results, one model for each water-quality category class is presented and discussed. From ecological point of view, each model is verified with the existing diatom indicator references found in literature by the biological expert.

Diatoms Classification with Weighted Averaging Fuzzy Operators for Eutrophication Prevention

The level of nutrients determines the triggering point of the eutrophication process, so it is very important to monitor these levels. Diatoms react rapidly on nutrient changes and that makes them ideal eutrophication bio-indicators. In the relevant literature there is known ecological reference for some diatoms, but for many of them these indicator features remain unidentified. In order to fill this gap and deal with disadvantages of the previous used methods, this research chapter aims to a novel fuzzy algorithm for classification of diatoms for eutrophication prevention. By using this algorithm, each diatom will be classified and based on results from the diatom models will be recommended for such objective. The proposed method uses sigmoid distribution to reveal the diatom-indictor relationship. Combined with weighted averaging fuzzy operators the experimental results have verified and discovered several diatom indicators that can be used for prevention. Once the diatom is found in the water sample, the expert looks up in the database and identifies the health state of the ecosystem. This also can be done for metal parameters, not just for nutrients.

Rule Induction of Physical-Chemical Water Property from Diatoms Community

In this paper we use the property of diatoms as bioindicators, to indentify which physical-chemical parameters are contained in the taken sample using machine learning algorithm – CN2. Important physical-chemical parameters such as conductivity, saturated oxygen, pH, organic chemical parameters and metals are important in the process of environmental monitoring. These physical-chemical parameters have influence on the entire lake web food chain, thus disturbing the organism’s patterns and interactions between them, such as diatoms community. These communities have high coefficient of indication on certain process such as eutrophication and presence or absence of certain physical-chemical parameters, which means that they can be used as bio-indicators of water quality. The machine learning algorithm – CN2 can produce rules in a form IF-THEN which is suitable for organizing knowledge from diatoms abundance data. In literature the diatoms have ecological preference organized in the same manner. The experimental setup is build to satisfy not only the algorithm properties, but also the ecological knowledge of the diatoms community. We used several modifications of the algorithm, from which then we compare the compactness and coverage of the induced rule. Nevertheless, for regression problems we compare the correlation coefficient, root mean square error (RMSE) and relative root mean square error (RRMSE) or rule quality to point which experiment proved to be most accuracy and more general. Several of the rules are presented in this paper together with the evaluation performance.

Data Management for the Water Monitoring System of Lake Prespa

One of the main goals of the TRABOREMA project was related to the design of the integrated database system. The created database system should provide a convenient, easy-to-use, intuitive way of storing the captured data. The integrated database system represents a centralized storage facility which enables the users to better organize, control, manage and use the data, create reports, perform statistical analysis, establish patterns in the model of the data, etc.