Feature selection for clustering using instance-based learning by exploring the nearest and farthest neighbors

Chen, Chien‐Hsing

doi:10.1016/j.ins.2015.05.019

Cited by 15 publications

(12 citation statements)

References 53 publications

(47 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After defining S i , the salient features are identified based on their ability to distinguish between the nearest and farthest neighbours. The j th feature s j is more salient for i th data point, if it is more dependent on the target variable z and it is less redundant in comparison with other features (Chen, ). Accordingly, the salience vector for the i th data point is defined as u i = [ u i 1 … u ij … u if ] where the salience of the j th feature is calculated according to Equation :

u_{ij} = D ((),, s_{j}, c) - R ((), s_{j}),

where D (.…”

Section: Methodsmentioning

confidence: 99%

“…In the last two equations, MI (. ) denotes mutual information criterion that is calculated for the two discrete random variables of X and Y (Chen, ).…”

Section: Methodsmentioning

confidence: 99%

“…Equation is used to minimize the loss to find the optimal value of w ( t opt ). In this equation, α ( t ) is a monotonically descending learning rate and its initial value is usually determined empirically and its subsequent values are consecutively reduced at each iteration based on Equation (Chen, ).

α ((), t) = α ((), t - 1) \times ((), \frac{t_{italicopt} - t}{t_{opt}}) .

…”

Section: Methodsmentioning

confidence: 99%

“…The jth feature s j is more salient for ith data point, if it is more dependent on the target variable z and it is less redundant in comparison with other features (Chen, 2015). Accordingly, the salience vector for the ith data point is defined as…”

Section: Feature Weighting and Selectionmentioning

confidence: 99%

“…Performance of the method is examined in combination with cluster analysis as one of the most widely used methods in the regionalization studies (e.g., Acreman & Sinclair, 1986;Basu & Srinivas, 2014, 2015Burn, 1989;Burn & Goel, 2000;Dodangeh, Soltani, Sarhadi, & Shiau, 2014;Hall & Minns, 1999;Hall, Minns, & Ashrafuzzaman, 2002;Jingyi & Hall, 2004;Kingston, Hannah, Lawler, & McGregor, 2011;Ramachandra Rao & Srinivas, 2006a, 2006bSrinivas, Tripathi, Ramachandra Rao, & Govindaraju, 2008;Toth, 2013;Wiltshire, 1986). Ramachandra Rao and Srinivas (2006b) indicated that some hybrid algorithm can improve results of clustering by combining the strengths of two categories of clustering algorithms, hierarchical and partitional algorithms.…”

mentioning

confidence: 99%

See 4 more Smart Citations

A feature weighting and selection method for improving the homogeneity of regions in regionalization of watersheds

Ahani

Nadoushani

Moridi

2018

Hydrological Processes

View full text Add to dashboard Cite

The method used for feature selection or feature weighting in regionalization of watersheds may affect the results of regionalization methods considerably. It can play a key role in forming hydrologically homogeneous regions for regional flood frequency analysis. In this study, a method based on exploring the nearest and farthest neighbours of data points is introduced for identifying salient features for regionalization of watersheds. The method includes options to relate watershed features to flood data records in order to increase the homogeneity of the regions. The nearest and farthest neighbours are identified based on the criteria such as the mutual information criterion and Spearman's rank correlation coefficient. Then, the watershed features more able to explain the relationships between the nearest and farthest neighbours are identified as salient features to form homogeneous features for regional flood frequency analysis. The results show that the optimum option of the proposed method improves the performances of the hard and fuzzy clustering algorithms in more than half of the cases based on the cluster validity indices. Furthermore, the results reveal that the optimum option can increase the number of the homogeneous regions formed by clustering algorithms to a great extent. By using the optimum option with 5 nearest and 5 farthest neighbours, longitude, drainage area, and run‐off coefficient are identified as the salient features to regionalize Sefidrud basin. The results show that the proposed method can be considered as an efficient method to form homogeneous regions for regional flood frequency analysis.

show abstract

u_{ij} = D ((),, s_{j}, c) - R ((), s_{j}),

where D (.…”

Section: Methodsmentioning

confidence: 99%

“…In the last two equations, MI (. ) denotes mutual information criterion that is calculated for the two discrete random variables of X and Y (Chen, ).…”

Section: Methodsmentioning

confidence: 99%