BSP TO HARNESS ‘BIG DATA’ TO IMPROVE ANALYTICS
The Bangko Sentral ng Pilipinas (BSP) expects to roll out by 2023 its News Sentiment Index (NSI) as the regulator forays into the use of “big data” and artificial intelligence to bridge an information gap as well as supplement existing consumer and business surveys.
BSP Governor Benjamin Diokno said last week the NSI would capture relevant views on key macroeconomic events that may affect the current and emerging economic and financial environment.
“The NSI will leverage on big data, machine learning, and artificial intelligence to enhance the BSP’s monitoring activities for policy development and macro-financial surveillance,” Diokno said.
The project involves the development of software that will automatically gather information on topics of interest, such as consumer and business sentiment, from online news sources.
“The information gathered will, in turn, be processed using algorithms to derive the general sentiment (positive, neutral and negative) on the economic and financial front,” the BSP chief said.
Citing studies of other central banks and researchers, Diokno said NSI results closely reflected those of the respondent-based sentiment surveys, such as the Business and Consumer Expectations Surveys which the BSP releases every quarter.
Asked how the NSI could address disinformation online, Diokno said the survey would cover only financial, business and economic news produced by “recognized and reputable media outlets in the Philippines—those that have garnered awards or citations and are recognized by other [media] organizations, either locally or internationally.”
“Of course, we must differentiate between the release of inaccurate information and ‘fake news,’” Diokno said.
“If a news outlet inadvertently publishes wrong data or information, this is not fake news,” he said. “It is an error that will eventually be corrected by way of an erratum or an updated version of the article.”
The BSP defines big data as information at high volume, velocity or variety that cannot be processed using conventional tools and software. They need specific technology and analytical algorithms to transform into value for mission-critical processes.
Diokno said the use of big data would further enhance the BSP’s economic research and analyses, formulation of monetary and financial stability policies, statistical compilation as well as financial regulation and supervision.