MACHINE LEARNING & ARTIFICIAL INTELLIGENCE
Machine learning is used for generating information and knowledge via a technical system, e.g., a software. The system learns patterns or structures using previous examples and is subsequently able to independently evaluate and classify information and data.
Examples and application scenarios for machine learning are:
Sentiment analysis, content monitoring, technology categorization, predictive coding, clustering, alerting, and documents search.
Machine learning has already become very important in the context of big data since it enables processing large amounts of data quickly and easily.
WHY MACHINE LEARNING?
The era of big data leads to a huge surge of digital information. Storage space is now relatively cheap and pretty much limitless. However, this leads to a new challenge:
What previously was reserved only for the human mind – to understand text contents and contexts – is impossible today without technical assistance due to the mass of information. Luckily, highly professional machine learning software solutions are here to help.
THE DIFFERENCE BETWEEN MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE (AI)
Machine learning is often mentioned in direct context with artificial intelligence. These two approaches are gaining particular importance in the big data sector.
While artificial intelligence rather refers to the broad approach of ‘intelligent’ machines, machine learning is generally associated with software-based approaches. Machine learning is therefore a subarea of artificial intelligence.
Instead of focusing on extensive, system-based computer work, machine learning tries to mimic human decision-making processes based on the patterns learned.
SCHEMATIC REPRESENTATION: MACHINE LEARNING FOR DOCUMENT CLASSIFICATION
Define the target categories, which the documents need to be categorized in (diagnoses, schemata, etc.).
Provide the documents that have already been categorized into classes.
3. Auto Categorization
Let the system classify more documents, including a statement on the security.
4. Results Checking
Check the results with the least security for optimum learning success.
Machine Learning vs. traditional Approach
Data / Rules
Data / Examples