A simple analytics use case is predicting mortgage default. We can use different predictive models, e.g., C5 classification model in SPSS or Spark ML Random Forest Classifier in the notebook.
Category: Machine Learning
Unraveling Layers of Unstructured Data
In today’s fast paced business environment often a combination of capabilities is needed to achieve tangible business benefit, to build solutions that gather content from multiple sources, traditional databases all the way to including web and social. With those solutions, you can store, analyze, and report on data by using a combination of analytics and data science capabilities to drive actionable insights and visualization.
Unstructured data over unstructured content usually refers to text, sound, pictures, video. The analysis of unstructured content or unstructured data starts with the notion of cognitive capture with the ability to OCR the content.
Pulling structured data out of the depths of unstructured content allows us to gain access to that content.
Amalgamation takes that derived content and looks it up in systems of record, to match it with a Master Data Management system, to match it with a customer profile, a market segment, a recommendation, a basket analysis, a collaborative filtering.
Entity Resolution
Context.
Different Entities may be named differently but might actually be the same Entity.
Problem
Identities are not unique. How can we tell if they are probably associated with the same Entity,
Forces
Each entity name has a unique identifier.
Solution
Market Basket Analysis using association rule learning is a candidate algorithm for performing account resolution analysis.
Consequences
You have probabilistic indication of convergence on entity from possible multiple entities
AI Patterns 