國立虎尾科技大學 |

The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets./
Author:	Ndifon, Victor Bah.
Description:	1 online resource (131 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-07, Section: B.
Contained By:	Dissertations Abstracts International85-07B.
Subject:	Information technology. -
Online resource:	click for full text (PQDT)
ISBN:	9798381220070

The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets.
Ndifon, Victor Bah.

The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets. - 1 online resource (131 pages)

Source: Dissertations Abstracts International, Volume: 85-07, Section: B.

Thesis (Ph.D.)--Northcentral University, 2023.

Includes bibliographical references

Missing data remains one of the frequent quality issues whose effect becomes overwhelming downstream on machine learning applications if not detected and replaced efficiently and reliably. The problem is compounded in sparse data sets wherein about 50% of the data is missing. More information is needed on using the current imputation methods on sparse data sets. The problem this study addresses is the need for existing imputation methods to replace missing data reliably and efficiently in sparse data sets. Sparse data is a problem for data analysts, statisticians, and researchers across all fields who spend much time determining the adequate imputation method for such a data set. This quantitative descriptive research study evaluated the reliability and efficiency of predictive mean matching, K-near neighbor, decision trees, and multiple imputations by chained equations in replacing missing data in sparse data sets. The regression theoretical framework, complemented by the item response theoretical framework, guided this study. These frameworks help determine the probability of success in imputing the missing values in a data set containing categorical variables or limiting the output within a range of 0 and 1 or -1 and 1. These frameworks help in modeling single and multiple variables. Data from the University of California's data repository was analyzed to study the efficiency and reliability of the four multiple imputation methods above using Cronbach's alpha and RMSE, respectively, to answer the research questions. The findings revealed that the K-near neighbor and decision trees demonstrated higher reliability in sparse data sets. The study also showed no difference in the efficiency of the four imputation methods on sparse data sets. The results imply that researchers and data stewards can better handle reliability and efficiency issues in sparse data sets using the proper imputation methods. Future research should focus on crediting more complex missing data mechanisms, like missing completely at random and equally moving from an imputation method's default setting to various variations.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798381220070Subjects--Topical Terms:

559429
Information technology.
Subjects--Index Terms:

DataIndex Terms--Genre/Form:

554714
Electronic books.

The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets.
LDR:03401ntm a22003617 4500 001 1143879
005 20240517105017.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798381220070
035 $a (MiAaPQ)AAI30690694
035 $a AAI30690694
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Ndifon, Victor Bah. $3 1468689
245 1 4 $a The Reliability and Efficiency of Replacing Missing Data in Sparse Data Sets.
264 0 $c 2023
300 $a 1 online resource (131 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-07, Section: B.
500 $a Advisor: Kabia, Milton.
502 $a Thesis (Ph.D.)--Northcentral University, 2023.
504 $a Includes bibliographical references
520 $a Missing data remains one of the frequent quality issues whose effect becomes overwhelming downstream on machine learning applications if not detected and replaced efficiently and reliably. The problem is compounded in sparse data sets wherein about 50% of the data is missing. More information is needed on using the current imputation methods on sparse data sets. The problem this study addresses is the need for existing imputation methods to replace missing data reliably and efficiently in sparse data sets. Sparse data is a problem for data analysts, statisticians, and researchers across all fields who spend much time determining the adequate imputation method for such a data set. This quantitative descriptive research study evaluated the reliability and efficiency of predictive mean matching, K-near neighbor, decision trees, and multiple imputations by chained equations in replacing missing data in sparse data sets. The regression theoretical framework, complemented by the item response theoretical framework, guided this study. These frameworks help determine the probability of success in imputing the missing values in a data set containing categorical variables or limiting the output within a range of 0 and 1 or -1 and 1. These frameworks help in modeling single and multiple variables. Data from the University of California's data repository was analyzed to study the efficiency and reliability of the four multiple imputation methods above using Cronbach's alpha and RMSE, respectively, to answer the research questions. The findings revealed that the K-near neighbor and decision trees demonstrated higher reliability in sparse data sets. The study also showed no difference in the efficiency of the four imputation methods on sparse data sets. The results imply that researchers and data stewards can better handle reliability and efficiency issues in sparse data sets using the proper imputation methods. Future research should focus on crediting more complex missing data mechanisms, like missing completely at random and equally moving from an imputation method's default setting to various variations.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Information technology. $3 559429
650 4 $a Statistics. $3 556824
653 $a Data
653 $a Sparse data sets
653 $a Efficiency
655 7 $a Electronic books. $2 local $3 554714
690 $a 0489
690 $a 0463
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Northcentral University. $b School of Technology. $3 1466151
773 0 $t Dissertations Abstracts International $g 85-07B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30690694 $z click for full text (PQDT)