A Comparative Assessment of Support Vector Machines, Probabilistic Neural Networks, and K-Nearest Neighbor Algorithms for Water Quality Classification

Modaresi, Fereshteh; Araghinejad, Shahab

doi:10.1007/s11269-014-0730-z

Cited by 94 publications

(45 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A SVM for regression analysis is accomplished by solving a convex optimization problem, more specifically a quadratic programming problem [42]. BPNN and SVM have demonstrated their abilities in dealing with complicated datasets in many fields like classification [43,44] and data mining [45,46].…”

Section: Deriving Lai and Agb Via Vismentioning

confidence: 99%

Dynamic Mapping of Rice Growth Parameters Using HJ-1 CCD Time Series Data

et al. 2016

View full text Add to dashboard Cite

Abstract:The high temporal resolution (4-day) charge-coupled device (CCD) cameras onboard small environment and disaster monitoring and forecasting satellites (HJ-1A/B) with 30 m spatial resolution and large swath (700 km) have substantially increased the availability of regional clear sky optical remote sensing data. For the application of dynamic mapping of rice growth parameters, leaf area index (LAI) and aboveground biomass (AGB) were considered as plant growth indicators. The HJ-1 CCD-derived vegetation indices (VIs) showed robust relationships with rice growth parameters. Cumulative VIs showed strong performance for the estimation of total dry AGB. The cross-validation coefficient of determination (R 2 CV ) was increased by using two machine learning methods, i.e., a back propagation neural network (BPNN) and a support vector machine (SVM) compared with traditional regression equations of LAI retrieval. The LAI inversion accuracy was further improved by dividing the rice growth period into before and after heading stages. This study demonstrated that continuous rice growth monitoring over time and space at field level can be implemented effectively with HJ-1 CCD 10-day composite data using a combination of proper VIs and regression models.

show abstract

Section: Deriving Lai and Agb Via Vismentioning

confidence: 99%

Dynamic Mapping of Rice Growth Parameters Using HJ-1 CCD Time Series Data

et al. 2016

View full text Add to dashboard Cite

show abstract

“…Spark MLlib has supported several algorithms that could be used for classification, clustering, or regression. As proposed by (Modaresi and Araghinejad, 2014), (Ladjal et al, 2016), (Jaloree et al, 2014), (Saghebian et al, 2014), we use two classification algorithms, where the results would be compared to choose the best one. Classification algorithms which would be used were Support Vector Machine and Decision Tree.…”

Section: Learning Processmentioning

confidence: 99%

“…If ( , ) is dataset with = 1, … , , where is the vector containing features, and ∈ {−1,1} is the class label related to . SVM solves the following primal problem (Modaresi and Araghinejad, 2014), (Ladjal et al, 2016), (Suykens and Vandewalle, 1999).…”

Section: Learning Processmentioning

confidence: 99%

See 1 more Smart Citation

Classification extension based on IoT-big data analytic for smart environment monitoring and analytic in real-time system

Arridha

Sukaridhoto

Pramadihanto

et al. 2017

IJSSC

View full text Add to dashboard Cite

Abstract:Monitoring water conditions in real-time is a critical mission to preserve the water ecosystem in maritime and archipelagic countries, such as Indonesia that is relying on the wealth of water resources. To integrate the water monitoring system into the big data technology for realtime analysis, we have engaged in the ongoing project named SEMAR (Smart Environment Monitoring and Analytic in Real-time system), which provides the IoT-Big Data platform for water monitoring. However, SEMAR does not have an analytical system yet. This paper proposes the analytical system for water quality classification using Pollution Index method, which is an extension of SEMAR. Besides, the communication protocol is updated from REST to MQTT. Furthermore, the real-time user interface is implemented for visualisation. The evaluations confirmed that the data analytic function adopting the linear SVM and Decision Tree algorithms achieves more than 90% for the estimation accuracy with 0.019075 for the MSE.

show abstract

“…Vugrin et al [11] distinguished events from the fluctuation caused by working conditions by using trajectory cluster through feature extraction and cluster using fitting polynomial of data. Modaresi and Araghinejad [12] evaluated three methods: Support Vector Machine, Probabilistic Neural Network, and KNN of water quality event detection, and compared their performance.…”

Section: Introductionmentioning

confidence: 99%

Contamination Event Detection Method Using Multi-Stations Temporal-Spatial Information Based on Bayesian Network in Water Distribution Systems

Xie

et al. 2017

Water

View full text Add to dashboard Cite

Abstract:As a core part of protecting water quality safety in water distribution systems, contamination event detection requires high accuracy. Previously, temporal analysis-based methods for single sensor stations have shown limited performance as they fail to consider spatial information. Besides, abundant historical data from multiple stations are still underexploited in causal relationship modelling. In this paper, a contamination event detection method is proposed, in which both temporal and spatial information from multi-stations in water distribution systems are used. The causal relationship between upstream and downstream stations is modelled by Bayesian Network, using the historical water quality data and hydraulic data. Then, the spatial abnormal probability for one station is obtained by comparing its current causal relationship with the established model. Meanwhile, temporal abnormal probability is obtained by conventional methods, such as an Autoregressive (AR) or threshold model for the same station. The integrated probability that is calculated employed temporal and spatial probabilities using Logistic Regression to determine the final detection result. The proposed method is tested over two networks and its detection performance is evaluated against results obtained from traditional methods using only temporal analysis. Results indicate that the proposed method shows higher accuracy due to its increased information from both temporal and spatial dimensions.

show abstract

A Comparative Assessment of Support Vector Machines, Probabilistic Neural Networks, and K-Nearest Neighbor Algorithms for Water Quality Classification

Cited by 94 publications

References 19 publications

Dynamic Mapping of Rice Growth Parameters Using HJ-1 CCD Time Series Data

Dynamic Mapping of Rice Growth Parameters Using HJ-1 CCD Time Series Data

Classification extension based on IoT-big data analytic for smart environment monitoring and analytic in real-time system

Contamination Event Detection Method Using Multi-Stations Temporal-Spatial Information Based on Bayesian Network in Water Distribution Systems

Contact Info

Product

Resources

About