인터넷 검색추세를 활용한 빅데이터 기반의 주식투자전략에 대한 연구

Together with soaring interest on Big Data, now there are vigorous reports that unearth various social values lyingunderneath those data from a number of application areas. Among those reports many are using such data as Internetsearch histories from Google site, social relationships from Facebook, and transactional or locational traces collectedfrom various ubiquitous devices. Many of those researches, however, are conducted based on the data sets that areaccumulated over the North American and European areas, which means that direct interpretation and applicationof social values exhibited by those researches to the other areas like Korea can be a disturbing task. This research has started from a validation study against Korean environment of the former paper which saysan investment strategy that exploits up and down of Google search volume on a carefully selected set of terms showshigh market performance. A huge difference between North American and Korean environment can be eye witnessedvia the distinction in profit rates that are exhibited by the corresponding set of search terms. Two sets of searchterms actually presented low correlation in their profit rates over two financial markets. Even in an experiment whichcompares the profit rates with two different investment periods with the same set of search terms showed no suchmeaningful result that outperforms the market average. With all these results, we cautiously conclude that establishingan investment strategy that exploits Internet search volume over a specified word set needs more conscious approach.