1. Data Mining and Unsupervised Learning
In this video David explores the three different types of Data Mining mentioned in the precious presentation: Anomaly Detection, Clustering, and Association Mining.
Introduction: in this presentation David introduces and discusses the relationship between search and data mining and the types of search which can be achieved through different machine learning algorithms to produce relevant results.
Anomaly detection: moving on from search and data mining Dave introduces means by which we can find or detect objects or events which are unlike the objects we already know about i.e. point, collective or contextual anomalies.
Association mining: from detecting anomalies with association mining we Dave introduces the concept of co-occurrence within a data set and the calculation of rule support, confidence and lift.
Clustering: in a further look at 'finding me things that are like one-another' Dave introduces the idea of how we may use clustering of data points within a 'feature space' and the different means by which this clustering may be identified to produce boundaries.
Summary: How do we find things that are not like other things, find things that are related or find things that are like each other?