Clustering approaches for data with missing values: Comparison and evaluation

Himmelspach, Ludmila; Conrad, Stefan

doi:10.1109/icdim.2010.5664691

Cited by 23 publications

(20 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…PDS The Partial Distance Strategy [10,15]. Calculate the sum of squared differences of the mutually known components and scale to the missing components:…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Distance estimation in numerical data sets with missing values

Eirola

Doquire

Verleysen

et al. 2013

Information Sciences

View full text Add to dashboard Cite

The possibility of missing or incomplete data is often ignored when describing statistical or machine learning methods, but as it is a common problem in practice, it is relevant to consider. A popular strategy is to fill in the missing values by imputation as a pre-processing step, but for many methods this is not necessary, and can yield sub-optimal results. Instead, appropriately estimating pairwise distances in a data set directly enables the use of any machine learning methods using nearest neighbours or otherwise based on distances between samples. In this paper, it is shown how directly estimating distances tends to result in more accurate results than calculating distances from an imputed data set, and an algorithm to calculate the estimated distances is presented. The theoretical framework operates under the assumption of a multivariate normal distribution, but the algorithm is shown to be robust to violations of this assumption. The focus is on numerical data with a considerable proportion of missing values, and simulated experiments are provided to show accurate performance on several data sets.

show abstract

“…PDS The Partial Distance Strategy [10,15]. Calculate the sum of squared differences of the mutually known components and scale to the missing components:…”

Section: Methodsmentioning

confidence: 99%

“…A simple and widely used method for estimating distance with missing values is the Partial Distance Strategy (PDS) [10,15]. In the PDS, an estimate for the squared distance is found by calculating the sum of squared differences of the mutually known components, and scaling the value proportionally to account for the missing values.…”

Section: Related Workmentioning

confidence: 99%

Distance estimation in numerical data sets with missing values

Eirola

Doquire

Verleysen

et al. 2013

Information Sciences

View full text Add to dashboard Cite

show abstract

“…Conrad and Himmelspach [1] concluded that the missing values are considered MCAR, if the absence of data does not rely on data values in the data matrix that are experimented. It is believed that missing data cause biased result when performing data mining tasks because missing data are still considered as valuable representative attributes with respect to the hidden information in data sets.…”

Section: Introductionmentioning

confidence: 99%

CKNNI: An Improved KNN-Based Missing Value Handling Technique

Jiang

Yang

2015

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In data mining field, experimental data sets are often incomplete due to the imperfect nature of real world situations. However, the incompleteness of data sets generally leads to biased outcomes. Thus, data completeness is one of the most essential challenges among data mining tasks. In order to achieve better outcome many researchers have explored various techniques to reduce data incompleteness, and some existing methods have been widely used in real world applications. This paper first discusses some existing representative missing data handling techniques with their advantages and drawbacks. Then a new improved KNN based algorithm, Class-Based K-clusters Nearest Neighbor Imputation (CKNNI) is proposed, which integrates K-means cluster algorithm and conventional KNN algorithm to impute missing values in data sets. By clustering instances in the same class with K-means algorithm, CKNNI method then applies KNN algorithm to select a closest neighbor from the set of centroids in resulted clusters, and missing values are imputed with the ones from corresponding variables in a selected neighbor. Finally, the comparison based on multiple data sets indicates that CKNNI has improved the performance of KNN imputation significantly on large data sets yet comparative to other superior missing value handling algorithms.

show abstract

“…Grouping data were available for 26 interviews. Usually free pile-sorting data are analysed using cluster analysis [Bernard, 2006]; however, these analyses are sensitive to missing values [Himmelspach and Conrad, 2010], which in our case were generated by the exclusion of non-recognized animals from each free pile-sorting task. A large number of clustering algorithms that can deal with missing values has been proposed [Kaufman and Rousseeuw, 2005], and research in this field remains active.…”

Section: Identification Test and Species Groupingsmentioning

confidence: 99%

“…Folk taxonomies are usually derived using pile sort data [López et al, 1997;Koster et al, 2010;Papworth et al, 2013], whereby participants are asked to place a set number of taxa into groups which they perceive as containing animals that are similar. One of the caveats of the most commonly used pile sort analyses is that they are sensitive to missing data [Himmelspach and Conrad, 2010], and therefore require the researcher to use only species that will be recognized by the majority of participants, or will ask participants to sort animals which they do not recognize. The latter may be particularly problematic if the folk taxonomy of the community is based upon non-morphological characters, such as diet or the time of day during which a particular animal is active.…”

Section: Introductionmentioning

confidence: 99%

Know Your Monkey: Identifying Primate Conservation Challenges in an Indigenous Kichwa Community Using an Ethnoprimatological Approach

Stafford¹,

Alarcon-Valenzuela²,

Patiño³

et al. 2016

IJFP

View full text Add to dashboard Cite

Increasing pressure on tropical forests is continually highlighting the need to find new solutions that mitigate the impact of human populations on biodiversity. However, developing solutions that can tackle the drivers of anthropogenic pressure, or at least take them into account, hinges upon building a good understanding of the culture and perceptions of local people. This study aims to provide an overview of the ethnoprimatology of an indigenous Kichwa community in the Ecuadorian Amazon that maintains a traditional lifestyle but also has good access to markets. We examine whether primates are seen as a distinctive group and their relative importance as sources of bushmeat and as household pets. Pile-sorting exercises revealed that although locals generally group members of the order Primates together, tree-dwelling non-primates including sloths, coatis, kinkajous and tamanduas are also frequently classified as ‘monkeys'. The perceived importance of primates to the forest and the community lay more in their potential as bushmeat, and only 1 respondent identified an ecological role for the group in terms of seed dispersal. Gaining a better understanding of local perceptions will allow for better-informed conservation decisions that are more aware of potential impacts and are more likely to gain community support.

show abstract

Clustering approaches for data with missing values: Comparison and evaluation

Cited by 23 publications

References 8 publications

Distance estimation in numerical data sets with missing values

Distance estimation in numerical data sets with missing values

CKNNI: An Improved KNN-Based Missing Value Handling Technique

Know Your Monkey: Identifying Primate Conservation Challenges in an Indigenous Kichwa Community Using an Ethnoprimatological Approach

Contact Info

Product

Resources

About