Watson Natural Language Understanding for advanced text analytics
By Susan C. Daffron | 4 minute read | April 6, 2017
Text and other unstructured content from sources like social media posts, articles and blogs is filled with insights that might help your business. The problem is that it’s difficult to parse unstructured text to see trends. For example, are people making positive or negative comments about my product since it was released?
You can’t easily quantify and understand social media because it’s a big mass of unstructured text. Because the information can’t be mined, in many cases the data just sits there, unusable. Your data has always contained business-critical answers for your business. The problem was you couldn’t easily find the answers through all the noise. But with sophisticated natural language processing (NLP) software, you can. NLP allows the user to extract key metadata from their text, including entities, relations, concepts, sentiment, and emotion. Now your program can understand the complexities of human language to uncover meaningful insights.
NLP without a PhD
In March 2017, IBM announced the evolution of AlchemyLanguage as a stand-alone service, rebranded as IBM Watson® Natural Language Understanding. This cloud-based Watson service offers a suite of NLP capabilities that make it possible to quickly and easily extract and analyze metadata from unstructured text.
For an introduction into why you should use Watson Natural Language Understanding, watch this video:
Watson Natural Language Understanding gives you access to cognitive technologies that used to only be available to expert data scientists. It extends customization options for specific domains and industries. With a few simple API calls, you can use Natural Language Understanding to automate NLP and machine learning capabilities within your applications. It provides a comprehensive NLP stack that scales easily as a cloud-based service.
Like AlchemyLanguage, but better
Natural Language Understanding builds on the legacy of its predecessor AlchemyLanguage. It remains a leader in text analytics and offers a few key improvements. Natural Language Understanding features the familiar AlchemyLanguage capabilities through a simplified API. It was re-architected to run natively on Bluemix and is software as a service (SaaS).
Natural language processing capabilities for advanced text analytics
Natural Language Understanding includes the core functions of AlchemyLanguage along with some improvements and consolidation. Here are a few key differentiating features:
Sentiment and emotion
Natural Language Understanding returns both overall sentiment and emotion for a document and targeted sentiment and emotion towards keywords in the text. For example, while a customer review may have an overall negative sentiment, particular keywords in the review may have a positive tone, which allows a deeper analysis of the text.
Using Natural Language Understanding, you can adapt entity and relationship extraction with custom models for specific uses. With IBM Watson Knowledge Studio, subject matter experts can collaborate on the creation of custom models without having to write a single line of code. These models can then be easily deployed to Natural Language Understanding to identify industry and domain-specific entities and relationships in unstructured text.
New pricing and free forever plan
Natural Language Understanding has improved pricing and a free forever plan. The pricing structure has been simplified. It is based on just the types of metadata that need to be extracted and the amount of data analyzed. The new pricing is particularly attractive if you’re deploying custom models from Watson Knowledge Studio. It’s easy to experiment with the service because you can deploy and use one custom model with the Natural Language Understanding free forever plan and because Watson Knowledge Studio has a free plan as well, you don’t have to pay anything to start customizing Natural Language Understanding.
You can easily track and manage your usage of Natural Language Understanding with the Bluemix usage dashboard. By setting up email spending notifications, you can be alerted to the spending thresholds you specify, so you can find out when you’ve reached 80%, 90% and 100% spending. You also can edit each spending notification as your needs change.
Natural Language Understanding has broad language coverage. It can understand text in nine different languages, including English, French, German, Spanish, Portuguese, Italian, Russian, Arabic and Swedish. You can customize Natural Language Understanding’s ability to detect entities and relationships with Watson Knowledge Studio in English, Arabic, Brazilian Portuguese, French, German, Italian, Japanese, Korean and Spanish. (IBM is adding more languages all the time, so check back soon if your favorite language isn’t supported today.)
What’s next with Natural Language Understanding?
If you’ve been using the IBM Watson AlchemyAPI for similar applications, you can get details about moving to Watson Natural Language Understanding in our documentation. Starting on April 7, 2017, it will no longer be possible to create a new instance of AlchemyAPI on Bluemix. Existing service instances will be supported until March 7, 2018. You can try out the Natural Language Understanding service on Bluemix and with the free plan, you receive 1000 free Enrichment items per day. If you want to try out customization with the free plan, you can use a Watson Knowledge Studio custom model.