Machine learning (ML) is about teaching computer programs to recognize images, words and sounds by giving them a lot of examples to learn from.
In very broad strokes, by creating machine learning algorithms, we aim to make a computer “see”, “hear” and “read”. For a computer to acquire such cognitive abilities, which are natural for our brain, it needs to be programmed to identify the patterns in various types of data and to compute what these patterns are about.
Once developed and trained, machine learning algorithms help create systems that can automatically respond to and interpret data. For example, a machine learning algorithm that has been exposed to your preferences data (what you like reading or listening to) would be able to compute these data with time and further predict what else you might like in order to serve it to you.
What is Machine Learning in the Context of Text Analysis?
In the field of text analysis, machine learning is indispensable for managing textual sources on a scale.
When it comes to text and its understanding, machine learning is best applied for automating and scaling processes such as:
- Relationships discovery in a free-flowing text (how is one document related to another or how is a person, who is mentioned in a text, related to another person who is mentioned in a different text).
- Annotation of textual resources (what does the text talk about and what are the people, events and locations in this text related to).
- Classification of documents (what is the topic of a given text and how is it related to other texts about similar topics).
The generic approach to teaching a machine to help us with the above-listed processes is:
- Giving the machine plenty of documents to learn from (in a format that the machine can read).
- Giving the machine context in the form of machine-readable reference (such as Wikidata or another knowledge base that the machine can use in order to link an entity to the relationships it enters with other entities).
How does Machine Learning Help Ontotext Build Solutions?
At Ontotext, machine learning is at the core of every solution related to semantic enrichment and content classification.
NOW or Never
Take for instance News on the Web (NOW) – Ontotext’s free public service, showcasing the opportunities Ontotext Platform opens up to Media and Publishing companies.
In order to identify mentions of people, places and companies in the news from around the Web and classify their content, NOW uses machine learning algorithms that are key to the automated processing of texts:
- Semantic Enrichment: a machine learning algorithm automatically links the mentions it has identified in the text to other entities in a knowledge base.
The resulting Knowledge Graph provides machines with concept and entity awareness about people, organizations and locations, and enhances data discovery.
- Classification: a machine learning algorithm automatically classifies the news it is fed with in categories and subcategories.
This classification of news provides automated context-sensitive analysis and helps users quickly access relevant information.
Learn more about NOW in our blog post: 4 Things NOW Lets You Do With Content.
Want to learn more about machine learning in the context of text analytics?
White Paper:
Knowledge Graphs in the Enterprise
