You may have heard about Aylien awhile back, when it was trying to carve a niche as a consumer products company that used its text analysis API to inform its delivery of articles via a news reader interface to the masses. It’s changed tactics since then: Now the company — which recently brought semantic web expert and lecturer at NUI Galway and Insight Centre Dr. John Breslin on-board as an advisor — is orienting its text analysis and news APIs to media and PR organizations, as well as other industries and developers.
“It makes a lot of sense for customers,” says CEO and founder Parsa Ghaffari. “The biggest volume of information on the Internet is represented as text, which is obviously unstructured information. Through NLP you can obviously try to find structure in text, …so we think it can be seen as a first step to the semantic web vision to enable any developer or startup to extract that structure and find the value in it.”
Aylien’s text analytics API consists of eight distinct natural language processing, information retrieval, and machine learning APIs for article extraction, article summarization, classification, entity extraction, concept extraction, language detection, sentiment analysis and hashtag suggestions.
“All eight are to extract insight and meaning from documents,” Ghaffari says, and the API is provided as a service. What sets Aylien apart from competitive solutions, including APIs from companies such as Alchemy and OpenCalais, he says, is that “we have the richest feature set with the most important NLP tasks in the API. It’s the most rounded package of functionality,” vs. other tools that might provide only entity extraction or sentiment analysis, for instance. Aylien also focused hard on high quality data for training its algorithms, he notes, and adds that its pricing on average is about 15 percent cheaper than competitors. (There also is a free educational plan for researchers and students in the NLP and semantic fields.)
Also in Aylien’s bailiwick is its News API, which launched this month. The News API is its Text Analysis API applied to real-time feeds of news from major news outlets. It can be used to syndicate news based on topics, source and such in a very easy way, he says, and includes the ability to measure the performance of each article in social media –likes, shares, re-tweets, and so on. “It’s a powerful filter on top of news that’s getting published from major news sources,” he says.
Room For All
While noting the features he considers to be his company’s competitive advantages, Ghaffari also thinks that the market for text analytics is bigger than ever, so that the pie is getting larger for all players. Use cases, he notes, include scenarios such as public relations firms using the text analysis technology to automatically determine which blogs and news web sites are the right outlets for their press releases. “We try to find similarities [between the releases and the outlets] and build an index for all the blogs they want to target, and find the most relevant journalist or author in those blogs or news web sites,” he says.
“So as a journalist you receive fewer less-relevant releases and it reduces the shotgun approach people have to issuing press releases on the PR end.”
Among businesses currently leveraging its API are Azuqua.com, which provides an enterprise platform for analyzing streaming data from services like Twotter. “Using our API they can take actions based on the sentiment of tweets – for example, automatically open a ticket if a lot of people are complaining about a product,” he says. “It’s all in real-time and has deep integration with our product.”
Future plans include adding to the Text Analysis API base: Users will get the ability to train their own classifiers on their own particular information, for example, and informality detection will become available. “You have a piece of text and can find out how well-written it is for grammar and punctuation,” he explains.