TUTCRIS - Tampereen teknillinen yliopisto

TUTCRIS

Steps on Multi-Target Prediction and Adaptability to Dynamic Input

Tutkimustuotos

Standard

Steps on Multi-Target Prediction and Adaptability to Dynamic Input. / Aho, Timo.

Tampere University of Technology, 2012. 97 s. (Tampere University of Technology. Publication; Vuosikerta 1015).

Tutkimustuotos

Harvard

Aho, T 2012, Steps on Multi-Target Prediction and Adaptability to Dynamic Input. Tampere University of Technology. Publication, Vuosikerta. 1015, Tampere University of Technology.

APA

Aho, T. (2012). Steps on Multi-Target Prediction and Adaptability to Dynamic Input. (Tampere University of Technology. Publication; Vuosikerta 1015). Tampere University of Technology.

Vancouver

Aho T. Steps on Multi-Target Prediction and Adaptability to Dynamic Input. Tampere University of Technology, 2012. 97 s. (Tampere University of Technology. Publication).

Author

Aho, Timo. / Steps on Multi-Target Prediction and Adaptability to Dynamic Input. Tampere University of Technology, 2012. 97 Sivumäärä (Tampere University of Technology. Publication).

Bibtex - Lataa

@book{c2e2e5281a4c41b5a6ee45824f20c325,
title = "Steps on Multi-Target Prediction and Adaptability to Dynamic Input",
abstract = "How should we react to dynamically changing inputs in various areas of computer science? This is one of the main questions we discuss in this thesis. The problem is present both in machine learning environment coping with massive amount of data available today and on the low level programming of computers. One of the hot topics currently in machine learning are so called ensemble methods. An ensemble model is a collection of multiple divergent, often simple, base models. The variance of base models has been shown to give clear benefit to the predictive power over using a single model. Not surprisingly, ensemble methods also give new possibilities for coping with dynamic online inputs; we can simply reweight the base models to adapt. However, in this thesis we are especially interested in ensemble methods in a specific framework. In many practical problems, we have multiple related attributes that need to be predicted. For example, predicting the growth of ora or biological composition of water are tasks that can be presented with multiple attributes that clearly relate to each other. Recently there has been some progress on methods that gain both smaller and more accurate overall models by making use of relations between the predicted attributes. In this thesis, we show that we can achieve both small and accurate models with a rule based ensemble method Fire. The method is extensively evaluated experimentally. We also pull some strings together by showing how similar problems have been solved in separate areas of computer science. In machine learning, the problem of dynamically changing input has been studied under a term of concept drift. Similarly in algorithm and data structure analysis a notion of locality of reference has been present for long. We introduce a general framework that covers both of the problems and brie y go through the work done on both of the areas. We hope that by giving pointers to a bridge over the gap between the fields, researchers in both areas could be able to pick up some fruits on the other side.",
author = "Timo Aho",
note = "Awarding institution:Tampereen teknillinen yliopisto - Tampere University of Technology<br/>Submitter:Submitted by Timo Aho (timo.aho@tut.fi) on 2011-12-21T19:11:00Z No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)<br/>Submitter:Approved for entry into archive by Kaisa Kulkki(kaisa.kulkki@tut.fi) on 2011-12-22T09:57:05Z (GMT) No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)<br/>Submitter:Made available in DSpace on 2011-12-22T09:57:05Z (GMT). No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)",
year = "2012",
month = "1",
day = "27",
language = "English",
isbn = "978-952-15-2725-8",
series = "Tampere University of Technology. Publication",
publisher = "Tampere University of Technology",

}

RIS (suitable for import to EndNote) - Lataa

TY - BOOK

T1 - Steps on Multi-Target Prediction and Adaptability to Dynamic Input

AU - Aho, Timo

N1 - Awarding institution:Tampereen teknillinen yliopisto - Tampere University of Technology<br/>Submitter:Submitted by Timo Aho (timo.aho@tut.fi) on 2011-12-21T19:11:00Z No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)<br/>Submitter:Approved for entry into archive by Kaisa Kulkki(kaisa.kulkki@tut.fi) on 2011-12-22T09:57:05Z (GMT) No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)<br/>Submitter:Made available in DSpace on 2011-12-22T09:57:05Z (GMT). No. of bitstreams: 1 aho.pdf: 3070116 bytes, checksum: bf7a7b85b60824a28b51eabf5ba7a287 (MD5)

PY - 2012/1/27

Y1 - 2012/1/27

N2 - How should we react to dynamically changing inputs in various areas of computer science? This is one of the main questions we discuss in this thesis. The problem is present both in machine learning environment coping with massive amount of data available today and on the low level programming of computers. One of the hot topics currently in machine learning are so called ensemble methods. An ensemble model is a collection of multiple divergent, often simple, base models. The variance of base models has been shown to give clear benefit to the predictive power over using a single model. Not surprisingly, ensemble methods also give new possibilities for coping with dynamic online inputs; we can simply reweight the base models to adapt. However, in this thesis we are especially interested in ensemble methods in a specific framework. In many practical problems, we have multiple related attributes that need to be predicted. For example, predicting the growth of ora or biological composition of water are tasks that can be presented with multiple attributes that clearly relate to each other. Recently there has been some progress on methods that gain both smaller and more accurate overall models by making use of relations between the predicted attributes. In this thesis, we show that we can achieve both small and accurate models with a rule based ensemble method Fire. The method is extensively evaluated experimentally. We also pull some strings together by showing how similar problems have been solved in separate areas of computer science. In machine learning, the problem of dynamically changing input has been studied under a term of concept drift. Similarly in algorithm and data structure analysis a notion of locality of reference has been present for long. We introduce a general framework that covers both of the problems and brie y go through the work done on both of the areas. We hope that by giving pointers to a bridge over the gap between the fields, researchers in both areas could be able to pick up some fruits on the other side.

AB - How should we react to dynamically changing inputs in various areas of computer science? This is one of the main questions we discuss in this thesis. The problem is present both in machine learning environment coping with massive amount of data available today and on the low level programming of computers. One of the hot topics currently in machine learning are so called ensemble methods. An ensemble model is a collection of multiple divergent, often simple, base models. The variance of base models has been shown to give clear benefit to the predictive power over using a single model. Not surprisingly, ensemble methods also give new possibilities for coping with dynamic online inputs; we can simply reweight the base models to adapt. However, in this thesis we are especially interested in ensemble methods in a specific framework. In many practical problems, we have multiple related attributes that need to be predicted. For example, predicting the growth of ora or biological composition of water are tasks that can be presented with multiple attributes that clearly relate to each other. Recently there has been some progress on methods that gain both smaller and more accurate overall models by making use of relations between the predicted attributes. In this thesis, we show that we can achieve both small and accurate models with a rule based ensemble method Fire. The method is extensively evaluated experimentally. We also pull some strings together by showing how similar problems have been solved in separate areas of computer science. In machine learning, the problem of dynamically changing input has been studied under a term of concept drift. Similarly in algorithm and data structure analysis a notion of locality of reference has been present for long. We introduce a general framework that covers both of the problems and brie y go through the work done on both of the areas. We hope that by giving pointers to a bridge over the gap between the fields, researchers in both areas could be able to pick up some fruits on the other side.

M3 - Doctoral thesis

SN - 978-952-15-2725-8

T3 - Tampere University of Technology. Publication

BT - Steps on Multi-Target Prediction and Adaptability to Dynamic Input

PB - Tampere University of Technology

ER -