Open-domain Event Extraction and Embedding for Natural Gas Market Prediction

Abstract

We propose an approach to predict the natural gas price in several days using historical price data and events extracted from news headlines. Most previous methods treats price as an extrapolatable time series, those analyze the relation between prices and news either trim their price data correspondingly to a public news dataset, manually annotate headlines or use off-the-shelf tools. In comparison to off-theshelf tools, our event extraction method detects not only the occurrence of phenomena but also the changes in attribution and characteristics from public sources. Instead of using sentence embedding as a feature, we use every word of the extracted events, encode and organize them before feeding to the learning models. Empirical results show favorable results, in terms of prediction performance, money saved and scalability.

Streamlining event extraction with a simplified annotation framework

“…Although the accuracy of the close-domain models is promising, most datasets are still limited to specific domains like medical data, historical documents, or specific types of news (Vanegas et al, 2015 ; Björne and Salakoski, 2018 ; Han et al, 2018 ). Thus, to extract a generic event from more generalized corpora, open-domain event extraction was developed (Chau et al, 2019 ; Liu et al, 2019 ). The early model considers the headline phrase as an event and disambiguates the events using Wordnet (Miller, 1995 ) and word sense disambiguation (Chau et al, 2019 ).…”

“…Thus, to extract a generic event from more generalized corpora, open-domain event extraction was developed (Chau et al, 2019 ; Liu et al, 2019 ). The early model considers the headline phrase as an event and disambiguates the events using Wordnet (Miller, 1995 ) and word sense disambiguation (Chau et al, 2019 ). This method leads to suboptimal performance as the arguments of an event are not necessarily positioned next to an event trigger keyword.…”

Event Extraction by Associating Event Types and Argument Roles

“…The purpose of event extraction is to capture the event types that we are interested in from many texts and show the essential arguments of events in a structured form Cao, Peng, Wu, Dou, Li, and Yu (2021). The event extraction task is divided into open-domain based event extraction (Chau, Esteves, and Lehmann (2019); Liu, Huang, and Zhang (2019); ) and schema-based event extraction (Huang et al (2016); Yang and Mitchell (2016); Ferguson et al (2018); Ahmad, Peng, and Chang (2021)) according to whether to construct event schema for extracting triggers and arguments.…”

“…Therefore, many scholars study open domain event extraction ; Wang and Lillis (2020)). The open domain event extraction method obtains topics by clustering or classification and then extracts event arguments from multiple texts related to the topic (Chau et al (2019)). At present, the mainstream method is to first cluster the text, and then extract keywords from each type of text as event arguments (Chau et al (2019)).…”

“…The open domain event extraction method obtains topics by clustering or classification and then extracts event arguments from multiple texts related to the topic (Chau et al (2019)). At present, the mainstream method is to first cluster the text, and then extract keywords from each type of text as event arguments (Chau et al (2019)). The open domain event extraction method does not need lots of data annotation work, and can not specify a specific event category (Mejri and Akaichi (2017)).…”