Recognition of Implicit Geographic Movement in Text

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2054–2063 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2054 Recognition of Implicit Geographic Movement in Text Scott Pezanowski, Prasenjit Mitra Information Sciences and Technology The Pennsylvania State University University Park, PA, USA 16802 {scottpez, pmitra}@psu.edu Abstract Analyzing the geographic movement of humans, animals, and other phenomena is a growing field of research. This research has benefited urban planning, logistics, animal migration understanding, and much more. Typically, the movement is captured as precise geographic coordinates and time stamps with Global Positioning Systems (GPS). Although some research uses computational techniques to take advantage of implicit movement in descriptions of route directions, hiking paths, and historical exploration routes, innovation would accelerate with a large and diverse corpus. We created a corpus of sentences labeled as describing geographic movement or not and including the type of entity moving. Creating this corpus proved difficult without any comparable corpora to start with, high human labeling costs, and since movement can at times be interpreted differently. To overcome these challenges, we developed an iterative process employing hand labeling, crowd voting for confirmation, and machine learning to predict more labels. By merging advances in word embeddings with traditional machine learning models and model ensembling, prediction accuracy is at an acceptable level to produce a large silver-standard corpus despite the small gold-standard corpus training set. Our corpus will likely benefit computational processing of geography in text and spatial cognition, in addition to detection of movement. Keywords: geographic movement, spatial cognition, text classification, crowdsourcing, corpus, machine learning, embeddings 1. Introduction The study of movement of humans, animals, and other en- tities throughout geographic space has a large and growing body of research (Dodge et al., 2012; Dodge, 2016b; Dodge et al., 2016; Dodge, 2016a; Gonz´alez et al., 2008; Huang, 2017; Soares Junior et al., 2017). Datasets exist that allow scientists to produce valuable knowledge about these move- ments to improve urban planning, better understand ani- mal migrations, and detect unusual movements of vessels. However, text sources that contain descriptions of move- ment also exist such as emails, social media posts, web doc- uments, written historical documents, and other sources, and these sources are currently underutilized. In fact, IBM recently estimated that 80% of the World’s data comes in the form of unstructured data (Schneider, 2016) and per- haps the biggest form of unstructured data is text. Although data produced by GPS is becoming more common, under- going the process of using GPS to collect large-scale data is still expensive for many use-cases and therefore commonly does not exist for many entity types. For text sources that are identified as containing descriptions of movement, they can be converted into a traditional geographic data format using Geographic Information Retrieval (GIR) which auto- matically detects place mentions in text and resolves those place mentions to the intended place in the World, thereby allowing for further analysis. In order to make better use of this …

Content truncated to 500 words