Self-organization and missing values in SOM and GTM

Vatanen, Tommi; Osmala, Maria; Raiko, Tapani; Lagus, Krista; Sysi-Aho, Marko; Orešič, Matej; Honkela, Timo; Lähdesmäki, Harri

doi:10.1016/j.neucom.2014.02.061

Cited by 106 publications

(54 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…All SOMs were generated using the Imputation SOM algorithm [31] and the SOM Toolbox v2.1 for MATLAB [35], used with MATLAB 2015b (Mathworks: Natick, MA, USA). Info on the heuristics the MATLAB SOM Toolbox uses to determine the number of map units and aspect ratio of the map are provided by the authors of the toolbox [35].…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Relating Water Quality and Age in Drinking Water Distribution Systems Using Self-Organising Maps

et al. 2016

View full text Add to dashboard Cite

Understanding and managing water quality in drinking water distribution system is essential for public health and wellbeing, but is challenging due to the number and complexity of interacting physical, chemical and biological processes occurring within vast, deteriorating pipe networks. In this paper we explore the application of Self Organising Map techniques to derive such understanding from international data sets, demonstrating how multivariate, non-linear techniques can be used to identify relationships that are not discernible using univariate and/or linear analysis methods for drinking water quality. The paper reports on how various microbial parameters correlated with modelled water ages and were influenced by water temperatures in three drinking water distribution systems.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Classical SOM training ignores vector dimensions corresponding to null input values when calculating all distance metrics. Alternatively, when updating reference vectors during the iterative process, null values can be substituted for expectations by utilising an Imputation SOM (used herein) as described in Vatanen et al [31].…”

Section: Som Analysismentioning

confidence: 99%

Relating Water Quality and Age in Drinking Water Distribution Systems Using Self-Organising Maps

et al. 2016

View full text Add to dashboard Cite

show abstract

“…SOM imputation was performed using the Imputation SOM algorithm described in Vatanen et al (2015) and implemented in the SOM Toolbox for Matlab [20]. Mean imputation was implemented in Matlab.…”

Section: Methodsmentioning

confidence: 99%

An Overview and Evaluation of Recent Machine Learning Imputation Methods Using Cardiac Imaging Data

Liu

Gopalakrishnan

2016

Preprint

View full text Add to dashboard Cite

Many clinical research datasets have a large percentage of missing values that directly impacts their usefulness in yielding high accuracy classifiers when used for training in supervised machine learning. While missing value imputation methods have been shown to work well with smaller percentages of missing values, their ability to impute sparse clinical research data can be problem specific. We previously attempted to learn quantitative guidelines for ordering cardiac magnetic resonance imaging during the evaluation for pediatric cardiomyopathy, but missing data significantly reduced our usable sample size. In this work, we sought to determine if increasing the usable sample size through imputation would allow us to learn better guidelines. We first review several machine learning methods for estimating missing data. Then, we apply four popular methods (mean imputation, decision tree, k-nearest neighbors, and self-organizing maps) to a clinical research dataset of pediatric patients undergoing evaluation for cardiomyopathy. Using Bayesian Rule Learning (BRL) to learn ruleset models, we compared the performance of imputation-augmented models versus unaugmented models. We found that all four imputationaugmented models performed similarly to unaugmented models. While imputation did not improve performance, it did provide evidence for the robustness of our learned models.

show abstract

“…The SOM clustering algorithm [54] consists of an input layer composed of m neurons and a two-dimensional planar array competition layer composed of a*b neurons. The SOM clustering process is as follows: First, a smaller weight is assigned to the connection weights between the m input neurons and the output neurons.…”

Section: Som Algorithmmentioning

confidence: 99%

Location Recommendation of Digital Signage Based on Multi-Source Information Fusion

Xie

Zhang

et al. 2018

Sustainability

View full text Add to dashboard Cite

Abstract:With the increasing amount of digital signage and the complexity of digital signage services, the problem of introducing precise location recommendation methods for digital signage should be solved by digital signage enterprises. This research aims to provide a sustainable location recommendation model that integrates the spatial characteristics of geographic locations and multi-source feature data to recommend locations for digital signage. We used the outdoor commercial digital signage within the Sixth Ring Road area in Beijing as an example and combined it with economic census, population census, average house prices, social network check-in data, and the centrality of traffic networks that have an impact on the sustainable development of the regional economy as research data. The result shows that the proposed method has higher precision and recall in location recommendation, which indicates that this method has a better recommendation effect. It can further improve the recommendation quality and the deployment of digital signage. By this method, we can optimize resource allocation and make the economics and resources sustainable. The digital signage recommendation results of the Beijing City Sixth Ring Road indicated that the areas suitable for digital signage were primarily distributed in Wangfujing, Financial Street, Beijing West Railway Station, and tourist attractions in the northwest direction of the Fifth Ring Road. The research of this paper not only provides a reference for the integration of geographical features and their related elements data in a location recommendation algorithm but also effectively improves the science of digital signage layout, prompting advertising efforts to advance precision, personalization, low carbonization, and sustainable development.

show abstract

Self-organization and missing values in SOM and GTM

Cited by 106 publications

References 16 publications

Relating Water Quality and Age in Drinking Water Distribution Systems Using Self-Organising Maps

Relating Water Quality and Age in Drinking Water Distribution Systems Using Self-Organising Maps

An Overview and Evaluation of Recent Machine Learning Imputation Methods Using Cardiac Imaging Data

Location Recommendation of Digital Signage Based on Multi-Source Information Fusion

Contact Info

Product

Resources

About