September 15, 2016
Briefly
- From the Yorkshire Ranter: simple explanations of Brexit that aren’t supported by the data
- From Cardiogram: using the Apple Watch to diagnose abnormal heart rhythms
- From MIT Technology Review, an analysis of emotional patterns in fiction. “We find a set of six core trajectories which form the building blocks of complex narratives” They don’t really cover the possibility that they find six just because that’s as many as they can align neatly with their current approach..
- A reminder from the last US elections: ‘momentum’ doesn’t mean anything with votes — it’s a metaphor, not an analogy.
- From Hilda Bastian: The quality of a research study is rarely uniformly good across all the things it studies (though it could be uniformly bogus)
- On diagnosing depression from Instagram photos “But they’ve buried the real story. The depression rate among adults in the United States is 6.7%. The depression rate among the crowdsourced workers who shared their photos is 41.2%” (Medium)
Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »
They say the found the story arcs with other approaches. “Importantly, we also find these emotional arcs using two other methods: As clusters in a hierarchical clustering using Ward’s algorithm and as clusters using unsupervised
machine learning.” They also say that “less than 10 modes are enough to reconstruct
the emotional arc to a degree of accuracy visible to the eye.” Isn’t that reasonably open-minded about the alternatives?
8 years ago
I’m not denying the story arcs, but that’s not what’s new. The novel claim is that there are exactly six. If you had a story arc that varied more rapidly than up-down-up, it would be increasingly harder to align the examples in time to get clusters. Also, presumably they’d get rarer. I’m saying I don’t think they can reliably distinguish a hard cutoff on the number, which is what they imply.
8 years ago