Applying Text Mining for Identifying Future Signals of Land Administration

Abstract

Companies and governmental agencies are increasingly seeking ways to explore emerging trends and issues that have the potential to shape up their future operational environments. This paper exploits text mining techniques for investigating future signals of the land administration sector. After a careful review of previous literature on the detection of future signals through text mining, we propose the use of topic models to enhance the interpretation of future signals. Findings of the study highlight the large spectrum of issues related to land interests and their recording, as nineteen future signal topics ranging from climate change mitigation and the use of satellite imagery for data collection to flexible standardization and participatory land consolidations are identified. Our analysis also shows that distinguishing weak signals from latent, well-known, and strong signals is challenging when using a predominantly automated process. Overall, this study summarizes the current discourses of the land administration domain and gives an indication of which topics are gaining momentum at present.

BEAM: A Network Topology Framework to Detect Weak Signals

“…He used these metrics to build the Keywords Emergence Map (KEM) and the Keywords Issue Map (KIM). However, Lee and Park [10] then Krigsholm and Riekkinen [11] raise two pitfalls: 1) the problem of uncertainty when the same keyword is at the limit of several quadrants or appears in different quadrants for both maps; 2) the problem of interpretation that occurs when there are several meanings for a given keyword. In addition, keywords related to weak signals are, in general, isolated terms and the absence of relationships and context limits the information in further interpretation.…”

New perspectives for data‐supported foresight: The hybrid AI‐expert approach

“…That is, due to the rapid development of computing capacity, algorithms to analyze patterns in unstructured data sources—such as natural language processing (NLP), topic‐modeling such as latent Dirichlet allocation (LDA), and deep learning methods based on artificial neural networks—are increasingly available (Daas & van der Doef, 2020; LeCun et al, 2015; Mühlroth & Grottke, 2020; Porter, 2019). Recently, a number of proposals have been made on how to mobilize this potential in foresight, especially for identifying early signals of emerging changes (e.g., Krigsholm & Riekkinen, 2019; Lee & Park, 2018; Mühlroth & Grottke, 2018; 2020). Insights from these new data sources can, potentially, show new perspectives on data‐supported foresight for experts, policy makers, and strategists and their respective decision‐making processes.…”

“…These data analyses are largely operating on structured data sources, such as patents and publications (see e.g., Abbas et al, 2014; Daim et al, 2006; Glänzel et al, 2004; Huang et al, 2014; Huang & Chang, 2014; Milanez et al, 2014; Small et al, 2017; Zhang et al, 2018). In addition, these data analyses have relied upon the mining of structured, clearly defined databases for environmental scanning, weak signal scanning, or the extrapolation of time‐series to forecast technology trajectories (see e.g., Bengisu & Nekhili, 2006; Krigsholm & Riekkinen, 2019; Martino, 2003; Small et al, 2014).…”

Detection of Event Precursors in Social Networks: A Graphlet-Based Method

“…A keyword that has low visibility and a low diffusion level is considered a weak signal. Other studies leaned on these two indicators by adding a context to a list of keywords and used, for example, topic modeling such as LDA (Latent Dirichlet Allocation) [14,15] and clustering algorithms such as k-Means or k-Medoids [16].…”