Whether you are a business owner or an employee of a company, you notice large amounts of unstructured data being generated every day pertaining to the business. Sources for this data can be internal, such as emails, instant messages, documents, and wikis, as well as external, such as social media, news, and blogs.
There must be times when you ask yourself — how can I make sense of all this data, and use it as an advantage for the business?
You can invest significant time and resources, form a team of Natural Language, Data Science, and Machine Learning engineers, and start tinkering with data using various algorithms.
Or, you can simply opt to use one of the cloud-based text analytics services. These services provide out-of-the-box language and machine learning models for comprehending the meaning of textual content.
IBM Watson Natural Language Understanding (NLU) stands out among such services as one of the most comprehensive text analytics platforms. It provides out-of-the-box models for nine language understanding features, with support for customization.
Forrester, one of the most influential research and advisory firms in the world, recently ranked IBM as the leader in AI-based text analytics platforms. As part of its research, Forrester reviewed Watson NLU along with its competitors. You can read more about it here and grab a free copy of the report here.
You can use the simple API of NLU to pass in text, HTML or URL of a website, and request NLU features. Here’s a list of the features, along with some example use cases:
Sentiment — Show the overall sentiment of a certain brand on Twitter and how it is trending over time.
Keywords — Show most important words in a document as a tag cloud or feed them to a text classification algorithm.
Entities — Extract persons, locations, companies, job titles etc. from documents and the associated sentiment for each of them in the document.
Relation — Find out how entities are related in the text and build a knowledge graph. As an example, for the sentence “John Doe works at IBM”, NLU returns a relation “employedBy” between entities “John Doe/Person” and “IBM/Organization”
Categories — Classify text from documents or web pages on 1000+ categories or sub-categories.
Emotion — Detect tone of an SMS message such as the degree of joy, anger, disgust, fear, and sadness.
Concepts — Tag documents with abstract concepts contained in them.
Semantic Roles — Understand the relationship between subject and object in news content (who is doing what?). As an example, for the sentence “IBM acquires Redhat”, NLU returns “IBM/subject”, “Redhat/object”, “acquire/action”
Webpage metadata — Extract title, author, publication date, most prominent image from a webpage.
NLU supports thirteen different languages. Automatic language detection is performed on the text to be processed before appropriate language models are applied. The language models are continuously updated. Take a look at all the significant accuracy improvements made just this year.
Available all over the world, all the time
NLU, along with other Watson services, is available in six different geographic regions: US-South, US-East, Germany-Frankfurt, Australia-Sydney, Japan-Tokyo, and UK-London.
You can pick a region closest to where your applications are hosted, reducing network latency times. Within each of these regions, NLU is deployed in Kubernetes clusters, with worker nodes distributed across three different zones (i.e., physical data centers), ensuring a fault tolerant service that is available 24x7 365 days a year.
Try it out for free
You can try the full capabilities of Watson Natural Language Understanding for free using the Lite plan. You may find the following resources helpful in getting started.
Thanks for reading and have a wonderful holiday!