Cluster Computing and MapReduce Lecture 4

1:10

Thank you :)

frankly, the ones she talked about are not too advanced algorithms, but very popularly accepted and practical, but we should learn like that rather than in the acadamic, where the old nerds always play the complex math symbols all the times, and add useless things merely because they feel that is formal

For a good explanation of canopy clustering, and why you might want to use such algorithms, check out the Wikipedia article on canopy clustering, and the McCallum et al. paper on the topic. Basically, anything involving information retrieval will make a really good case for clustering algorithms. For example, string matching a million names against another set with a million names, where many of them do not match exactly, but would be considered equivalent.

she is hot, i want to date her.

MapReduce is really cool...but I think the future is combining structured and unstructured data by integrating with scalable databases...that's how you can really popularize analytics - check out Aster Data for example, they just announced a MapReduce database.

pathetic nerds trying to save their stupid BJ queen from disgrace... the fact is she got hired only because of her looks, since she obviously is a brainless twat.

She obviously understands the material, but cannot explain it in terms of practical application. I understood everything she said, but was constantly trying to apply it to the real world while she was talking. This is in stark contrast to the rest of the lectures in this series.

I desagree with you super.