Next Big Thing After Page Rank
As we learned from the lecture, Google paved a way for the new era of trusted search results, thanks to its sophisticated PageRank algorithm. It works by giving each webpage a value based on how many other websites link to it, based on the authorities of those other websites. As google succeeded in numerically analyzing authority of a website, which is, in essence, a human-created contents and thus hard to deduce its credibility, Cornell might be the next game changer in the arena of credibility analysis on the network.
Recently, it has been revealed that Amazon filed a huge lawsuit against fake reviewers who have been paid to post phony reviews, five dollars per review allegedly. Along with Google, Amazon is another major web network where billions of people visit on a daily basis to make easy purchases. People who may have limitations in the geographical setting, now have the privilege of enjoying a wide array of products from all around the world based on Amazon’s search results that automatically display highly ranked (from review section) products up top. It has become handy for those who often finds it hard to make a purchase decisions online. Not only does the search algorithm becomes affected by reviews, but also general customers do get affected greatly by the reviews for they are the warrants the true worth of a product, something that is hard to tell on the internet. Therefore, it has been the case that positive reviews push customers strongly to lead final transaction. Given such important role that customer review plays on the e-commerce industry, the media report on the rampant fake reviews on Amazon has been scandalous.
In the light of the growing e-commerce market and its significance to related businesses, Cornell researcher’s recent milestone in developing a software tool that detects fakeness of reviews is comparable to Google’s invention of PageRank. According to the article, this research team trained the software to identify a certain pattern in unknown fake reviews. When it was implemented to analyze 320 reviews from TripAdvisor, it detected that fake reviewers are prone to “use more verbs and less punctuation, and they focus more on family activities than the actual hotels”. The correctness was ninety percent.
As covered in the class, web search is unique in its own right because scale, dynamic, heterogeneity, reactivity exist. I believe efforts like Cornell’s fake detecting algorithm can ultimately evolve to reflect all of these natures of web search and become a trusted tool to protect customers from untrustworthy reviews so as to lead them to a right decision. It is really interesting to note how web search technologies and ideas advance in accordance with numerous variability caused my people of unknown wills.
http://www.pcworld.com/article/236655/Cornell_Software_Learns_How_to_Spot_Fake_Online_Reviews.html
http://fortune.com/2015/10/21/amazon-fake-product-reviews/