12 Nov Spotting New Flowers and Pulling Weeds… Online
Sunday’s record-setting sales of Alibaba Singles Day event at $30.8B reminded consumer brands, again, the importance of understanding online trends.
When Yimian Data use natural language processing to help clients understand e-commerce, there are two ongoing challenges: new words and strawman comments.
- Spotting new words: new words or new usages of existing words continuously emerge on e-commerce every week. It’s time-consuming to manually identify these new words, label them, and enter them into our libraries. Not done correctly, our tools will fail to discover emerging trends on the market, such as the new word “Singing like a Thunder”（歌雷呐）, the transliteration of the American explosive chewing gum brand “Grenades Gum”.
- Detect fake comments: online vendors frequently use humans or robots to batch generate fake comments. We call these “flooding comments” (水军）in Chinese, the equivalent of “strawman comments” in English. The flood of strawman comments was used to either promote someone’s own products or to attack competitors. If not extracted, these comments will skew our quantitative analyses of consumer concerns and sentiments.
Fortunately, Yimian team have been long using internal tools to discover new words and remove strawman comments.
Last week, our artificial intelligent team just released new versions of “New Word Discovery” tool and “e-Commerce Strawman Comments Detection” tool. The new releases included enhanced capabilities of new words filtering by the role of speech, new words ranking by importance in a consumer category, and improved accuracy in strawman detection.
These tools help our clients spot new trends earlier, and stay away from the plague of fake reviews. If they were gardeners, they can now do better in spotting new flowers and pulling weeds.
To learn more about Yimian Data services, please contact Rong Zhang.