Enhancing the applicability of Kohonen Self-Organizing Map (KSOM) estimator for gap-filling in hydrometeorological timeseries data

Nanda, Trushnamayee; Sahoo, Bhabagrahi; Chatterjee, Chandranath

doi:10.1016/j.jhydrol.2017.03.072

Cited by 17 publications

(11 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although SOM has been successfully used to classify the different metals in many research studies [40][41][42][43], the spatial assessments of water quality were first attempted to use SOM in the Jinghe Oasis in Xinjiang. Previous studies [44] found that the clustering results between conventional clustering analysis and SOM were generally similar.…”

Section: Discussionmentioning

confidence: 99%

Recognizing the Relationship between Spatial Patterns in Water Quality and Land-Use/Cover Types: A Case Study of the Jinghe Oasis in Xinjiang, China

Fei

Wang

2018

Water

View full text Add to dashboard Cite

To understand the relationship between spatial water quality patterns and changes in land-use/cover types in the Jinghe Oasis, 47 water sampling sites measured in May and October 2015 were divided into six cluster layers using the self-organizing map method, which is based on non-hierarchical k-means classification. The water quality indices included the chemical oxygen demand (COD), biological oxygen demand (BOD), suspended solids (SS), total phosphorus (TP), total nitrogen (TN), ammonia nitrogen (NH 3 -N), chromaticity (SD), and turbidity (NUT). Data was also collected on the changes in the farmland, forest-grassland, water body, salinized land, and other land types during the wet and dry seasons. Then, we combined these data with the classification results of the GF-1 remote sensing satellite data obtained in May and October 2015 and analyzed the influences of land-use/cover type on water quality for different layers and seasons. The results indicate that Clusters 1 to 3 included monitoring samples from the wet season (May 2015), whereas Clusters 4 to 6 included monitoring samples from the dry season (October 2015). In general, the COD, SS, NUT, TN, and NH 3 -N values were high around the Ganjia Lake Haloxylon natural conservation area in the southern Ebinur Lake region, east of Ebinur Lake, and around the Kuitun River during the wet season. The SD values around these areas were high. Moreover, high BOD and TP values were mainly concentrated around the Ganjia Lake Haloxylon natural conservation area, as well as the Kuitun River, during the dry season. In the discussion on the relationship between the different water quality parameters and land-use/cover type changes, we determined that farmland, forest-grassland, and salinized land significantly influenced the water quality parameters in the Jinghe Oasis. In addition, the influences of various land-use/cover types on the water quality parameters in the research zone during the different seasons exhibited the following descending order of magnitude: farmland → forest-grassland → salinized land → water body → others. Moreover, their influences were lower during the wet season than the dry season.In conclusion, developing research on the relationship between the spatial framework of the water quality in the Jinghe Oasis and land-use/cover type changes is significant for the time sequence distribution of water quality in arid regions from both theoretical and practical perspectives.

show abstract

Section: Discussionmentioning

confidence: 99%

Recognizing the Relationship between Spatial Patterns in Water Quality and Land-Use/Cover Types: A Case Study of the Jinghe Oasis in Xinjiang, China

Fei

Wang

2018

Water

View full text Add to dashboard Cite

show abstract

“…In addition, all clusters in a SOM map are neighboring [22]. Nanda et al used SOM for hydrological analysis [23]; Chu et al used SOM for their climate study [24]; Voutilainen et al clustered a gerontological medical dataset using SOM [25]; Kanzaki et al used SOM in their radiation study to analyze the liver damage from radon, X-rays, or alcohol treatments in mice [26]; and Tsai et al have clustered a dataset about water and fish species in an ecohydrological environment study [27].…”

Section: Related Workmentioning

confidence: 99%

“…Let us assume that C is obtained by Algorithm 1, Ω computed by C is the set of the clusters of the instances in I, I′ is a new set of the instances using a new attribute as the target (by filling it with Ω), IG is obtained by information gain feature selection method with I′, expressed in (23), fc is the highest ranked feature in IG, and FC is the set of the values in the fcth feature in C as in (14); sc is the second highest ranked feature in IG, and SC is the set of the values in the scth feature in C, as shown in the following equations:…”

Section: K-means++ and Mapping According To The Best Two Features Detmentioning

confidence: 99%

Novel Two-Dimensional Visualization Approaches for Multivariate Centroids of Clustering Algorithms

Doğan

Dalkılıç

Birant

et al. 2018

Scientific Programming

View full text Add to dashboard Cite

The dimensionality reduction and visualization problems associated with multivariate centroids obtained by clustering algorithms are addressed in this paper. Two approaches are used in the literature for the solution of such problems, specifically, the self-organizing map (SOM) approach and mapping selected two features manually (MS2Fs). In addition, principle component analysis (PCA) was evaluated as a component for solving this problem on supervised datasets. Each of these traditional approaches has drawbacks: if SOM runs with a small map size, all centroids are located contiguously rather than at their original distances according to the high-dimensional structure; MS2Fs is not an efficient method because it does not take features outside of the method into account, and lastly, PCA is a supervised method and loses the most valuable feature. In this study, five novel hybrid approaches were proposed to eliminate these drawbacks by using the quantum genetic algorithm (QGA) method and four feature selection methods, Pearson’s correlation, gain ratio, information gain, and relief methods. Experimental results demonstrate that, for 14 datasets of different sizes, the prediction accuracy of the proposed weighted clustering approaches is higher than the traditional K-means++ clustering approach. Furthermore, the proposed approach combined with K-means++ and QGA shows the most efficient placements of the centroids on a two-dimensional map for all the test datasets.

show abstract

“…The machine learning process used in this step can be categorized into two such as supervised learning and unsupervised learning based processes depending on data availability. Artificial Neural Network (ANN), Decision Tree (DT), Support Vector Machine (SVM), etc, are some of the supervised machine learning processes, and Self‐Organizing Maps (SOM), K‐means clustering, etc, are some of the unsupervised machine learning processes . Numerous works are already done in the case of disease categorization and forecast, but the existing methodologies are not able to sustain high performance for a number of datasets.…”

Section: Introductionmentioning

confidence: 99%

“…Artificial Neural Network (ANN), Decision Tree (DT), Support Vector Machine (SVM), etc, are some of the supervised machine learning processes, and Self-Organizing Maps (SOM), K-means clustering, etc, are some of the unsupervised machine learning processes. 10 Numerous works are already done in the case of disease categorization and forecast, but the existing methodologies are not able to sustain high performance for a number of datasets. The proposed methodology, ie, EFRC, focuses toward building a sustainable and effective disease categorization and foreseeing method with the ability to handle the missing attributes of data as well as the effective prediction of numerous disease obtained from various dataset with several categories of diseases.…”

Section: Introductionmentioning

confidence: 99%

Clustering‐based approach for medical data classification

Kodabagi

Tikotikar

2018

Concurrency and Computation

View full text Add to dashboard Cite

Summary Medical data records are growing enormously in present days due to the growth of various medical technologies and population of the globe. It is a difficult task for a medical expert to process these data records to identify and provide treatment regarding the disease of patient. Hence, it is necessary to automate the processing of such medical data using machine learning‐based Medical Decision Support System. In this proposed work, an Effective Fuzzy Rule Classifier (EFRC)–based decision support system is used for analysis of UCI medical dataset for identification of disease of patients. Initially, the method determines the best centroid value for each of the attribute using training dataset. Furthermore, the centroid value is refined using test samples. The refined centroid values of all the attributes are used by the fuzzy classifier for medical data classification. Experimental results have proven that EFRC performs better classification than existing systems in terms of accuracy, sensitivity, and specificity.

show abstract

Enhancing the applicability of Kohonen Self-Organizing Map (KSOM) estimator for gap-filling in hydrometeorological timeseries data

Cited by 17 publications

References 16 publications

Recognizing the Relationship between Spatial Patterns in Water Quality and Land-Use/Cover Types: A Case Study of the Jinghe Oasis in Xinjiang, China

Recognizing the Relationship between Spatial Patterns in Water Quality and Land-Use/Cover Types: A Case Study of the Jinghe Oasis in Xinjiang, China

Novel Two-Dimensional Visualization Approaches for Multivariate Centroids of Clustering Algorithms

Clustering‐based approach for medical data classification

Contact Info

Product

Resources

About