Normalized conformal prediction for timeseries data

Kowalczewski, Jakub (2018). Normalized conformal prediction for timeseries data. Thesis (Master thesis), E.T.S. de Ingenieros Informáticos (UPM).

Description

Title: Normalized conformal prediction for timeseries data
Author/s:
  • Kowalczewski, Jakub
Contributor/s:
  • Boström, Henrik
  • Girdzijauskas, Šarūnas
Item Type: Thesis (Master thesis)
Masters title: Data Science
Date: 2018
Subjects:
Freetext Keywords: Conformal prediction; Normalized conformal prediction; Time series; Forecasting
Faculty: E.T.S. de Ingenieros Informáticos (UPM)
Department: Otro
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

[img]
Preview
PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (455kB) | Preview

Abstract

Every forecast is valid only if proper prediction intervals are stated. Currentlymodels focus mainly on point forecast and neglect the area of prediction intervals.The estimation of the error of the model is made and is applied to everyprediction in the same way, whereas we could identify that every case is differentand different error measure should be applied to every instance. One ofthe state-of-the-art techniques which can address this behaviour is conformalprediction with its variant of normalized conformal prediction. In this thesiswe apply this technique into time series problems. The special focus is putto examine the technique of estimating the difficulty of every instance usingthe error of neighbouring instances. This thesis describes the entire process ofadjusting time series data into normalized conformal prediction framework andthe comparison with other techniques will be made. The final results do not showthat aforementioned method is superior over an existing techniques - in varioussetups different method performed the best. However, it is similar in terms ofperformance. Therefore, it is an interesting add-on to data science forecastingtoolkit.

More information

Item ID: 57817
DC Identifier: http://oa.upm.es/57817/
OAI Identifier: oai:oa.upm.es:57817
Deposited by: Biblioteca Facultad de Informatica
Deposited on: 31 Jan 2020 09:21
Last Modified: 31 Jan 2020 09:21
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM