Handling missing data from heteroskedastic and nonstationary data

Fulufhelo V. Nelwamondo; Tshilidzi Marwala

doi:10.1007/978-3-540-72383-7_151

Handling missing data from heteroskedastic and nonstationary data

Fulufhelo V. Nelwamondo, Tshilidzi Marwala

University of the Witwatersrand

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

8 Citations (Scopus)

Abstract

This paper presents a computational intelligence approach for predicting missing data in the presence of concept drift using an ensemble of multi-layered feed forward neural networks. An algorithm that detects concept drift by measuring heteroskedasticity is proposed. Six instances prior to the occurrence of missing data are used to approximate the missing values. The algorithm is applied to simulated time series data sets resembling non-stationary data from a sensor. Results show that the prediction of missing data in non-stationary time series data is possible but is still a challenge. For one test, up to 78% of the data could be predicted within 10% tolerance range of accuracy.

Original language	English
Title of host publication	Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings
Publisher	Springer Verlag
Pages	1293-1302
Number of pages	10
Edition	PART 1
ISBN (Print)	9783540723820
DOIs	https://doi.org/10.1007/978-3-540-72383-7_151
Publication status	Published - 2007
Externally published	Yes
Event	4th International Symposium on Neural Networks, ISNN 2007 - Nanjing, China Duration: 3 Jun 2007 → 7 Jun 2007

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	4491 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	4th International Symposium on Neural Networks, ISNN 2007
Country/Territory	China
City	Nanjing
Period	3/06/07 → 7/06/07

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-540-72383-7_151

Cite this

Nelwamondo, F. V., & Marwala, T. (2007). Handling missing data from heteroskedastic and nonstationary data. In Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings (PART 1 ed., pp. 1293-1302). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4491 LNCS, No. PART 1). Springer Verlag. https://doi.org/10.1007/978-3-540-72383-7_151

Nelwamondo, Fulufhelo V. ; Marwala, Tshilidzi. / Handling missing data from heteroskedastic and nonstationary data. Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings. PART 1. ed. Springer Verlag, 2007. pp. 1293-1302 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{c74d77a21c344e0999adff4afd5b5227,

title = "Handling missing data from heteroskedastic and nonstationary data",

abstract = "This paper presents a computational intelligence approach for predicting missing data in the presence of concept drift using an ensemble of multi-layered feed forward neural networks. An algorithm that detects concept drift by measuring heteroskedasticity is proposed. Six instances prior to the occurrence of missing data are used to approximate the missing values. The algorithm is applied to simulated time series data sets resembling non-stationary data from a sensor. Results show that the prediction of missing data in non-stationary time series data is possible but is still a challenge. For one test, up to 78% of the data could be predicted within 10% tolerance range of accuracy.",

author = "Nelwamondo, {Fulufhelo V.} and Tshilidzi Marwala",

year = "2007",

doi = "10.1007/978-3-540-72383-7_151",

language = "English",

isbn = "9783540723820",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

number = "PART 1",

pages = "1293--1302",

booktitle = "Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings",

address = "Germany",

edition = "PART 1",

note = "4th International Symposium on Neural Networks, ISNN 2007 ; Conference date: 03-06-2007 Through 07-06-2007",

}

Nelwamondo, FV & Marwala, T 2007, Handling missing data from heteroskedastic and nonstationary data. in Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 4491 LNCS, Springer Verlag, pp. 1293-1302, 4th International Symposium on Neural Networks, ISNN 2007, Nanjing, China, 3/06/07. https://doi.org/10.1007/978-3-540-72383-7_151

Handling missing data from heteroskedastic and nonstationary data. / Nelwamondo, Fulufhelo V.; Marwala, Tshilidzi.
Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings. PART 1. ed. Springer Verlag, 2007. p. 1293-1302 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4491 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Handling missing data from heteroskedastic and nonstationary data

AU - Nelwamondo, Fulufhelo V.

AU - Marwala, Tshilidzi

PY - 2007

Y1 - 2007

N2 - This paper presents a computational intelligence approach for predicting missing data in the presence of concept drift using an ensemble of multi-layered feed forward neural networks. An algorithm that detects concept drift by measuring heteroskedasticity is proposed. Six instances prior to the occurrence of missing data are used to approximate the missing values. The algorithm is applied to simulated time series data sets resembling non-stationary data from a sensor. Results show that the prediction of missing data in non-stationary time series data is possible but is still a challenge. For one test, up to 78% of the data could be predicted within 10% tolerance range of accuracy.

AB - This paper presents a computational intelligence approach for predicting missing data in the presence of concept drift using an ensemble of multi-layered feed forward neural networks. An algorithm that detects concept drift by measuring heteroskedasticity is proposed. Six instances prior to the occurrence of missing data are used to approximate the missing values. The algorithm is applied to simulated time series data sets resembling non-stationary data from a sensor. Results show that the prediction of missing data in non-stationary time series data is possible but is still a challenge. For one test, up to 78% of the data could be predicted within 10% tolerance range of accuracy.

UR - http://www.scopus.com/inward/record.url?scp=37249075926&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-72383-7_151

DO - 10.1007/978-3-540-72383-7_151

M3 - Conference contribution

AN - SCOPUS:37249075926

SN - 9783540723820

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1293

EP - 1302

BT - Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings

PB - Springer Verlag

T2 - 4th International Symposium on Neural Networks, ISNN 2007

Y2 - 3 June 2007 through 7 June 2007

ER -

Nelwamondo FV, Marwala T. Handling missing data from heteroskedastic and nonstationary data. In Advances in Neural Networks - ISNN 2007 - 4th International Symposium on Neural Networks, ISNN 2007, Proceedings. PART 1 ed. Springer Verlag. 2007. p. 1293-1302. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-540-72383-7_151

Handling missing data from heteroskedastic and nonstationary data

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this