Graph-Based Clustering Approach for Economic and Financial Event Detection Using News Analytics Data

Sidorov, Sergei; Faizliev, Alexey; Levshunov, Michael; Chekmareva, Alfia; Gudkov, Alexander A.; Korobov, Eugene

doi:10.1007/978-3-030-01159-8_26

Cited by 1 publication

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When composing a graph, some studies use full articles as graph nodes (Sidorov et al 2018), while others use keywords extracted defined via TF-IDF scoring (Sayyadi, Hurst, and Maykov 2009), a supervised keyword labelling (Liu et al 2017), named entities (Moutidis and Williams 2019), and word groups (Liu et al 2020). Our method utilizes a knowledge graph-based linguistic model that has been specifically extended to include concepts from the medical domain.…”

Section: Related Workmentioning

confidence: 99%

“…Further, events can be determined by extracting graph communities (Moutidis and Williams 2019;Sidorov et al 2018), by iteratively removing edges with high "betweenness" scores (Liu et al 2017), by depth search for graph connectivity components (Sidorov et al 2018). We use the Infomap community detection algorithm (Rosvall, Axelsson, and Bergstrom 2009) that allows moving iteratively from coarse to fine community structure in the network.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Identifying Early Warning Signals from News Using Network Community Detection

Vine

Boxer

Dinani

et al. 2022

AAAI

View full text Add to dashboard Cite

The paper addresses the challenge of accelerating identification of changes in risk drivers in the insurance industry. Specifically, the work presents a method to identify significant news events ("signals") from batches of news data to inform Life & Health insurance decisions. Signals are defined as events that are relevant to a tracked risk driver, widely discussed in multiple news outlets, contain novel information and affect stakeholders. The method converts unstructured data (news articles) into a sequence of keywords by employing a linguistic knowledge graph-based model. Then, for each time window, the method forms a graph with extracted keywords as nodes and draws weighted edges based on keyword co-occurrences in articles. Lastly, events are derived in an unsupervised way as graph communities and scored for the requirements of a signal: relevance, novelty and virality. The methodology is illustrated for a Life & Health topic using news articles from Dow Jones DNA proprietary data set, and assessed against baselines on a publicly available news data set. The method is implemented as an analytics engine in Early Warning System deployed at Swiss Re for the last 1.5 years to extract relevant events from live news data. We present the system's architectural design in production and discuss its use and impact.

show abstract

Section: Related Workmentioning

confidence: 99%