UNDERSTAND THE GLOBAL ECONOMIC CRISIS: A TEXT SUMMARIZATION APPROACH

• Author: Benoit Favre,Shuhua Liu
• Published date: 01 April 2013
• Date: 01 April 2013
• DOI: http://doi.org/10.1002/isaf.1340

UNDERSTAND THE GLOBAL ECONOMIC CRISIS:

A TEXT SUMMARIZATION APPROACH

SHUHUA LIU

a

*AND BENOIT FAVRE

b

a

Arcada University of Applied Sciences, Department of Business, ITand Media, Jan-Magnus Janssonin aukio 1, 00550 Helsinki,

Finland

b

LIF, Aix-Marseille Université, Parc Scientifique et Technologique de Luminy, 163, avenue de Luminy –Case 901, F-13288

Marseille Cedex 9, France

SUMMARY

Economic crises are significant threats to macroeconomic stability. They can incur large costs and bring devastating

effects on economies, with the effects often spilling over into other economies. Since 2007we have witnessed the most

severe and widely spread economic crisis since the Great Depression of the 1930s. In the meantime, a huge amount of

ongoing media coverage, reporting, analysis and debate concerning the global economic crisis has been generated. In

this study we explore the possibilities of applying text summarization tools to learn from text documents the various

discussions surrounding the global economic crisis. Included in our analysis are blog posts and articles of highly influ-

ential economists, as well as official speeches and publications of government organizations. The ICSI-ILP extractive

summarizer is applied in a large number of experiments, and the summary outputs are manually examined and evalu-

ated. The results provide us with insights into the potential and limitations of state-of-the-art summarization systems

when used to help us quickly learn and digest large amounts of textual information. The results also suggest different

ways to break the limitations of text summarization technology. Copyright © 2013 John Wiley & Sons, Ltd.

Keywords: text summarization; content extraction; text analysis; economic crisis

1. INTRODUCTION

The world economy of today has evolved into an extensive interdependent network through

globalized trade, investment, financial markets, supply chains, commodity flows and people flows. This

greatly increases dynamics and introduces elements of instability into economic systems. Since 2007

we havewitnessed the worst global economiccrisis since the Great Depression of the 1930s.The subprime

mortgage crisis that started in the USA in 2007 turned into a heated credit crunch in 2008 and became a

global recession in 2009. It reduced global real activity and trade to a degree unprecedented since World

War II, contributed to the failure of big businesses and significant decline in econo mic activity and

necessitated substantial financial commitments from governments (Cecchetti et al., 2009). As of 2010,

after more than 3years of turbulent movements and digestions, the crisis seemed to be under control,

and significant risks for theworld economy seemedto be fading away. However, a newwave of turbulence

in the form of debt crises has brought new threats to the stabilityof the world economy.

1

* Correspondence to: Shuhua Liu,Arcada University of Applied Sciences, Department of Business, IT and Media, Jan-Magnus

Janssonin aukio 1, 00550 Helsinki, Finland. E-mail: Shuhua.liu@arcada.fi

1

As of October 2012, Nouriel Roubini is predicting that the global economy is again on course for a ‘perfect storm’in 2013.

Copyright © 2013 John Wiley & Sons, Ltd.

INTELLIGENT SYSTEMS IN ACCOUNTING, FINANCE AND MANAGEMENT

Intell. Sys. Acc. Fin. Mgmt. 20, 89–110 (2013)

Published online in Wiley Online Library (wileyonlinelibrary.com)DOI: 10.1002/isaf.1340

In the course of the world experiencing one of the most severe and widely spread economic crises

in history, huge amounts of writings and discussions concerning various aspects of the crisis have

been produced in the past few years in newswire, social media, research institutions and government

organizations. We now have easy online access to large amounts of information-rich and opinion-rich

accounts and debates of what has happened in the economic world. Is it possible and how can we make

use of existing computational technology tools to explore such abundant data so that one can quickly

develop a sufficient overall understanding of the crisis and can lear n about key economic issues related

to the crisis? What would be the problems and limitations of the technologies? To answer these

questions, in this study we draw upon developments in methods and techniques for natural language

processing, especially in text summarization, to analyse a large collection of text documents concerning

the global economic crisis. Included in our analysis are selected blog posts, commentary columns and

articles by highly influential economists, together with speeches, reports and publications of governmen-

tal and international organizations.

Natural languagetechnology has experiencedfast and extensive developments in the past twodecades.

Advancements in the field of artificial intelligence, especially in machine learning methods and ontology

development efforts, fuelled by the easy access to large amounts of textual information on the web and

active research in information retrieval and search engineshave contributed to the development in compu-

tational linguistics and text analytics. Awide variety of useful methodsand tools have flourished, tackling

issues in topicdetection and tracking, information extraction,text summarization, questionanswering and,

most recently, sentimentanalysis and opinionmining from text documents. Thesedevelopments are shown

in the explosion in the number of publications in computational linguistics, natural language processing,

information retrieval, text mining and text analytics conferences.

Text summarization is a process of distilling the most important content from text documents. The

first text summarization methods were invented by Luhn (1999) and Edmundson (1999). Very active

and intensive research efforts from the computational linguistics community are seen since the

1990s. Much progress has been made in exploring a variety of text summarizati on methods and tech-

niques (Paice, 1990; Salton et al., 1994; Mani and Maybury, 1999; Hovy and Lin, 1999; McKeown

and Radev, 1999; Erkan and Radev, 2004; Filatova and Hatzivassiloglou, 2004; Radev et al., 2004).

In most recent years, research on summarization continues in the direction of incor porating more

and more progress made in computational linguistics/natural language processing, domain-specific

ontology development efforts, advanced machine learning methods and summary evaluation methods

(Hennig et al., 2008; Li et al., 2008; Otterbacher et al., 2008; Ouyang et al., 2010; Wei et al., 2010).

In the meantime, in addition to news summarization, application research started to appear, for

example, in the summarization of emails, product reviews, medical dialogues, multilingual or

multimodal sources of varying types on the Web, such as blogs, and talk-show transcriptions

(the DARPA-funded GALE project; Galley, 2006).

Different from most of the research work reported in the literature that deals with controlled environ-

ment and laboratory problems, in this study we apply text analysis and summarization methods to a

real-life problem. This raises various questions that we will try to address in this paper:

1. Data. What sources should be covered and how can the data be collected?

2. Summary output. What would be a proper length constraint for the summaries and how can the

results be presented to the user in an easier to understand and useful way?

3. The system. How does a state-of-the-art generic system perform on a real-life topic? How can we

adapt it to process large amounts of data or to be deployed in interactive scenarios? What is the

impact of retrieval noise/off-topic data on summarization?

90 S. LIU AND B. FAVRE

Copyright © 2013 John Wiley & Sons, Ltd. Intell. Sys. Acc. Fin. Mgmt. 20,89–110 (2013)

DOI: 10.1002/isaf