Keywords is our keyword extractor that identifies the "information carrying" words text.

Using free Japanese text as input, Keywords analyses and extracts the most important words in the text while leaving out uninteresting words.

Example Newspaper article is courtesy of Asahi Shimbun

Try hovering to highlight words in the below newspaper article excerpt.

囲碁ガール新風 、 碁会所チェンジ 「 脳 からキレイに」

紫煙 くゆらす オヤジ が 占拠 する 碁会所 のイメージが変わりつつある。主役は「 囲碁ガール 」。「 脳 からキレイに」「 女子力アップ 」。 うたい文句 にいざなわれ、女性が続々と碁の世界へ。彼女たちから未来の名人が輩出するかも?
囲碁ガール新風 碁会所チェンジ からキレイに」

紫煙 くゆらす オヤジ 占拠 する 碁会所 のイメージが変わりつつある。主役は「 囲碁ガール 」。「 からキレイに」「 女子力アップ 」。 うたい文句 にいざなわれ、女性が続々と碁の世界へ。彼女たちから未来の名人が輩出するかも?

The important words are highlighted in green, and their relative importance is indicated by colour intensity.

Application areas

Keywords supports a range of important applications within search and analytics:

  • Social media analytics. Understand the gist or dominant trend of what people are saying in social media about a given brand, company or topic across a large document set.
  • Searching by example. Instead of inputting a few terms that describe what you are looking for, with Keywords you can provide sample passages of text (or whole documents) and have a large query automatically built to match the provided content.

    Negative examples are also possible. For example, give me documents about "apple", but I'm not interested in fruit so include some fruit articles as negative examples to make sure "fruity" articles will not be included in my search result.

    These queries tend to be too hard for people to formulate themselves.

  • Content based personalisation. Many personalisation approaches today are based on so-called collaborative filtering, i.e. personalisation based on events around an item or document, but doesn't take the content of the document itself into account. Keywords can be used to make content-based personalisation and recommendation systems.

Try right here, right now

We have made Keywords available here to give you an idea of what it does.

Tip Try inputting the body text of a Japanese newspaper article. Also try the various result formats.

Feature summary

  • High quality keyword extraction. Keywords extracts high quality keywords and phrases from input text. Extracted keywords are weighted and ranked according to their importance. Positions in the source documents to the extracted terms are also provided.
  • Easy of use and easy integration. Keywords is provided both as a Java SDK and as a REST/JSON server. This allows you to integrate Keywords with your application in the manner that is best for you.
  • Scalable and failure-resilient architecture. Keywords is easily deployable in a scalable and failure-resilient manner using off-the-shelf hardware and software load-balancers.
  • Active vocabulary. Keywords understands the current vocabulary of your document set and does not have a hardwired language model. As a result, it can pick up new terminology as you provide documents and it can also forget terminology as documents are deleted.

Frequently Asked Questions

Where can I learn more about Keywords?

Please contact us at moc.akilita@selas and we will be glad to give you more information about Keywords.

How much does Keywords cost?

Please contact us at the above e-mail address for a quote.

Can I receive a trial or evaluation license?

Most certainly. Please drop us a line and we will arrange a free trial license and provide download instructions.

Can you provide Keywords as a hosted service?

Yes, we can. Please contact us at the above e-mail address for a quote.

Which languages does Keywords support?

Keywords is designed to support Japanese, but Western languages can also be supported.

How can I use Keywords to do personalisation or query-by-example?

Please contact us and we will tell you more about the details for how to do this.