Скачать или смотреть Big Data Opportunity: Structured vs. Unstructured Data

http://www.patrickschwerdtfeger.com/sbi/

Where's the opportunity in Big Data? Is it with structured data or unstructured data? Experts estimate that over 95% of the data in the world today is unstructured and only 5% is structured, so there's definitely a lot MORE unstructured data to be mined. The case histories so far suggest that the biggest opportunities lie in the messy unstructured data; the data the INCLUDES the outliers rather than marginalize them. The outliers add the most interesting insights to the process and allow the algorithms to calculate probabilities using the entire sample size, rather than relying on sampling inferences based on a small subset of the population. So research your unstructured data. Look at all those machine logs and metadata and see what insights you might be able to glean. Those are the building blocks for predictive analytics and algorithms that value.