A dynamic programming approach to missing data estimation using neural networks

Nelwamondo, Fulufhelo V.; Golding, Dan; Marwala, Tshilidzi

doi:10.1016/j.ins.2009.10.008

Cited by 68 publications

(24 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NELWAMANDO employed data mining to a data set of information concerning HIV-patients such as age, education level, region etc. The imputation process is based on neural networks [28].…”

Section: State Of the Artmentioning

confidence: 99%

Improving Data Consistency in Production Control by Adaptation of Data Mining Algorithms

et al. 2016

View full text Add to dashboard Cite

Manufacturing companies are increasingly exposed to volatile market conditions. In this environment, ensuring a reliable adherence to promised delivery dates, allows for a considerable competitive advantage. However, due to dynamically changing production circumstances and high varieties in production programs, manufacturing companies regularly fail in reaching this logistical target. A main prerequisite for mastering this challenge are excellent Production Planning and Control processes. The quality of transactional data of production processes are a commonly ignored root cause for inadequate detailed scheduling plans although a vast volume of these data are used for updating production job statuses and short-term production plans, deriving conclusions for immediate control interventions as well as monitoring production efficiency. Typically, measures for improving data quality involve implementing integrity constraints in databases and setting up data quality processes as well as dedicated organizational structures. Evidently, these classic approaches do not successfully prevent manufacturing companies from dealing with inadequate data quality in their PPC processes. Consequently, this paper presents a model for increasing the quality of data relevant for production processes by adapting data mining algorithms. This new approach allows to estimate probable values for typical data inconsistencies in transactional data of PPC processes. Several adapted algorithms are benchmarked on real-world data sets of German mid-sized manufacturing companies and evaluated towards their power and efficiency.

show abstract

“…NELWAMANDO employed data mining to a data set of information concerning HIV-patients such as age, education level, region etc. The imputation process is based on neural networks [28].…”

Section: State Of the Artmentioning

confidence: 99%

Improving Data Consistency in Production Control by Adaptation of Data Mining Algorithms

et al. 2016

View full text Add to dashboard Cite

show abstract

“…In [21] , a novel technique for missing data estimation using a combination of dynamic programming, neural networks and genetic algorithms (GA) on suitable subsets of the input data was presented, which is well suitable for decision making processes. The online and offline data imputation models were proposed based on the auto associative neural networks in [25] .…”

Section: Missing Data Imputationmentioning

confidence: 99%

Data imputation for gas flow data in steel industry based on non-equal-length granules correlation coefficient

Zhao

Liu

et al. 2016

Information Sciences

View full text Add to dashboard Cite

“…Other approaches such as model-based methods [17] and expectation maximization (EM) [18] may have higher accuracy, but are computationally expensive and problem specific. In clustering based single imputation (SI) methods [19, 20], data are first clustered using the non-missing values and then missing values are imputed using the instances of the cluster that contain the missing value instance.…”

Section: Related Workmentioning

confidence: 99%

Combining Fourier and lagged k -nearest neighbor imputation for biomedical time series data

Rahman

Huang

Claassen

et al. 2015

Journal of Biomedical Informatics

View full text Add to dashboard Cite

Most clinical and biomedical data contain missing values. A patient’s record may be split across multiple institutions, devices may fail, and sensors may not be worn at all times. While these missing values are often ignored, this can lead to bias and error when the data are mined. Further, the data are not simply missing at random. Instead the measurement of a variable such as blood glucose may depend on its prior values as well as that of other variables. These dependencies exist across time as well, but current methods have yet to incorporate these temporal relationships as well as multiple types of missingness. To address this, we propose an imputation method (FLk-NN) that incorporates time lagged correlations both within and across variables by combining two imputation methods, based on an extension to k-NN and the Fourier transform. This enables imputation of missing values even when all data at a time point is missing and when there are different types of missingness both within and across variables. In comparison to other approaches on three biological datasets (simulated and actual Type 1 diabetes datasets, and multi-modality neurological ICU monitoring) the proposed method has the highest imputation accuracy. This was true for up to half the data being missing and when consecutive missing values are a significant fraction of the overall time series length.

show abstract

A dynamic programming approach to missing data estimation using neural networks

Cited by 68 publications

References 8 publications

Improving Data Consistency in Production Control by Adaptation of Data Mining Algorithms

Improving Data Consistency in Production Control by Adaptation of Data Mining Algorithms

Data imputation for gas flow data in steel industry based on non-equal-length granules correlation coefficient

Combining Fourier and lagged k -nearest neighbor imputation for biomedical time series data

Contact Info

Product

Resources

About