take a look at makes use of text-mining to improve marketplace intelligence on startups

"enterprise giants like Google, Microsoft and Yahoo are spending tens of billions of bucks a year on acquiring smaller corporations for marketplace front, strategic highbrow belongings and proficient employees, but face a real task figuring out businesses with the right merchandise or technology inside the significant startup universe," said Gene Moo Lee, UTA assistant professor of information structures and Operations management.

"Our new approach uses massive data analytics and a text-mining approach known as subject matter modeling to pick out capacity fits," Lee introduced. "by means of studying unstructured, publicly available descriptions of any startups' business, we can quantify any  companies' business, geographic, investor and social proximity and from there identify ability objectives for mergers and acquisitions."

The researchers have tested the applicability in their studies with the aid of developing a cloud-based totally facts system based totally on their method and feature even launched a brand new business enterprise, topic technologies, that makes use of the machine to provide marketplace intelligence services on competitors, investors, acquisition objectives and capability enterprise companions to organizations and startups across the high-generation sector.

Mary Whiteside, meantime chair of information systems and Operations control inside UTA's university of enterprise, emphasized that this studies paperwork an essential part of UTA's strategic consciousness on facts-pushed discovery within the Strategic Plan 2020: bold solutions international effect.

"This studies demonstrates the capability transformation massive records analytics can convey to enterprise intelligence with using external facts assets and text mining Whiteside said. "topic modelling offers entrepreneurs, undertaking capitalists and analysts a brand new manner to navigate the continuously converting landscape of mergers and acquisitions."

For the preliminary evaluation, the researchers used publicly available statistics from startup database CrunchBase on 24,382 groups, the extensive majority of which were privately held, early-degree startups. For every organization, they took into account the headquarter location, enterprise quarter, cofounders, board contributors, key employees, investments, and the enterprise description, which was typically restrained to at least one or extra paragraphs on the important thing facts approximately the companies' merchandise, markets and technologies.

They then hired the textual content mining approach called subject matter modeling, which analyzes the language used within the startups' enterprise descriptions round shared products, technologies and markets. The startups' commercial enterprise proximity changed into then quantified based totally at the similarity of these topic descriptions.

The probability of a likely merger among two organizations became then computed deliberating business proximity, geographic vicinity, social hyperlinks between people in the two corporations and commonplace investor ownership, reflecting the strongly networked nature of the startup international.

"This records-driven, analytics-primarily based technique has proved powerful in explaining mergers and acquisitions in the startup world and complements existing toolkits for measuring enterprise proximity," Lee said. "Our gadget is particularly appropriate when the firms underneath observe are small and privately held so enterprise type in largely unavailable, that's the case for startups."

Lee and his co-researchers Zhan (Michael) Shi, assistant professor of records structures on the W.P. Carey college of enterprise at Arizona nation college and Andrew Whinston, Hugh Cullen Chair Professor on the McCombs college of commercial enterprise at the university of Texas at Austin, are also publishing their studies in the main journal control information structures Quarterly as "towards a better measure of business Proximity: topic Modeling for industry Intelligence." The paper is imminent however presently available as a preprint.

similarly to his instructional research, Lee has giant industry revel in at Samsung Electronics, AT&T Labs, Intel and Goldman Sachs. His studies hobbies encompass big-scale records analytics with packages on cell ecosystems, social network analysis and net protection. He holds ten patents in cellular era.

Shi's studies pursuits mirror the interface of economics and computation, with packages in social media, on line markets and innovation. His research has been published in numerous top academic journals and conferences.

Whinston is the Hugh Cullen Chair Professor inside the statistics, danger and Operation management department on the McCombs school of commercial enterprise at the university of Texas at Austin. He is likewise director on the center for research in electronic commerce.