University of Twente Student Theses


Automatic Weak Signal Detection and Forecasting

Gutsche, T. (2018) Automatic Weak Signal Detection and Forecasting.

[img] PDF
Abstract:Analyzing foresight approaches combined with weak signal mining reveals three significant shortcomings. First, foresight approaches are built on snapshot data, qualitative methods and lack of automation. All of which is insufficient in a constantly changing environment. Second, previous scholars, which combine foresight with weak signals focus on the sole detection of weak signals. Third, neither detecting weak signals, nor waiting for them to become strong signals is sufficient for strategic decision making. This study addresses these shortcomings with and aims to support strategic issue management by automatically detecting weak signals and forecasting their appearance through temporal web mining and time-series analysis. The usefulness of the study and its results are exemplified by a case study analyzing the environment of web conferencing solutions. The results show that the appearance of weak signals can be forecasted with a high f1-score and practitioners value the resulting system in terms of usability and utility.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Business Administration MSc (60644)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page