statistics   201471

« earlier    

Domestic Violence: The Secret Killer That Costs $8.3 Billion Annually
In the U.S., 24 percent of adult women and 14 percent of adult men have been physically assaulted by a partner at some point in their lives. It is the most common cause of injury for women ages 18 to 44. And it leads to an increased incidence of chronic disease: Abused women are 70 percent more likely to have heart disease, 80 percent more likely to experience a stroke and 60 percent more likely to develop asthma.

Nearly a quarter of employed women report that domestic violence has affected their work performance at some point in their lives. Each year, an estimated 8 million days of paid work is lost in the U.S. because of domestic violence.
domestic_violence  statistics  $ 
18 hours ago by Quercki
seaborn: statistical data visualization
Seaborn is a Python visualization library based on matplotlib. It provides a high-level interface for drawing attractive statistical graphics.
statistics  visualization  programming  library  pydata  python 
21 hours ago by newtonapple
Statistical skepticism: How to use significance tests effectively: 7 challenges & how to respond to them | Error Statistics Philosophy
Here are my slides from the ASA Symposium on Statistical Inference : “A World Beyond p < .05” in the session, “What are the best uses for P-values?”. (Aside from me,our session included Yoav Benjamini and David Robinson, with chair: Nalini Ravishanker.)

7 QUESTIONS

Why use a tool that infers from a single (arbitrary) P-value that pertains to a statistical hypothesis H0 to a research claim H*?
Why use an incompatible hybrid (of Fisher and N-P)?
Why apply a method that uses error probabilities, the sampling distribution, researcher “intentions” and violates the likelihood principle (LP)? You should condition on the data.
Why use methods that overstate evidence against a null hypothesis?
Why do you use a method that presupposes the underlying statistical model?
Why use a measure that doesn’t report effect sizes?
Why do you use a method that doesn’t provide posterior probabilities (in hypotheses)?
p-values  statistics 
23 hours ago by deprecated
Statistical skepticism: How to use significance tests effectively: 7 challenges & how to respond to them
"""
1. Why use a tool that infers from a single (arbitrary) P-value that pertains to a statistical hypothesis H0 to a research claim H*?
2. Why use an incompatible hybrid (of Fisher and N-P)?
3. Why apply a method that uses error probabilities, the sampling distribution, researcher “intentions” and violates the likelihood principle (LP)? You should condition on the data.
4. Why use methods that overstate evidence against a null hypothesis?
5. Why do you use a method that presupposes the underlying statistical model?
6. Why use a measure that doesn’t report effect sizes?
7. Why do you use a method that doesn’t provide posterior probabilities (in hypotheses)?
"""
statistics  significance 
yesterday by aapl
How the Shape of a Weakly Informative Prior Affects Inferences
"""
When building sophisticated models, principled priors that quantify principled prior information are critical to ensuring reasonable inferences. Weakly informative priors are particularly well-suited to this task, identifying natural scales in the problem to regularize the information extracted from the likelihood. The exact shape of those priors, however, has important practical consequences.

Lighter tailed priors, such as a Gaussian distribution, induce strong regularization towards the given scales. If the scales are consistent with the measurements then this regularization yields well-behaved posteriors that facilitate statistical computation. If the scales are poorly-chosen, however, then the regularization biases the resulting inferences as would any poor assumption in the model.

Heavier tailed priors induce weaker regularization which allows the posterior to assign significant probability mass above the chosen scale where the computation of the model may become slow or even unstable. On the other hand, the weaker regularization does not bias the model as strongly when the scales are inconsistent with the measurements.

Because accurate and fast statistical computation is critical to any analysis, we on the Stan Development Team have come to favor the lighter tails of a weakly informative prior with a Gaussian shape as the most robust default. The scale of that prior then becomes another assumption that must be carefully analyzed in order the verify the accuracy of the resulting inferences.
"""
bayes  statistics 
yesterday by aapl
A Primer on Bayesian Multilevel Modeling using PyStan
"Multilevel models are regression models in which the constituent model parameters are given probability models. This implies that model parameters are allowed to vary by group. Observational units are often naturally clustered. Clustering induces dependence between observations, despite random sampling of clusters and random sampling within clusters."
bayes  statistics  multilevel  stan 
yesterday by aapl
Modelling Loss Curves in Insurance with RStan
"Loss curves are a standard actuarial technique for helping insurance companies assess the amount of reserve capital they need to keep on hand to cover claims from a line of business. Claims made and reported for a given accounting period are tracked seperately over time. This enables the use of historical patterns of claim development to predict expected total claims for newer policies."
bayes  statistics  stan 
yesterday by aapl
Splines In Stan
"We start by providing a brief introduction to splines and then explain how they can be implemented in Stan. We also discuss a novel prior that alleviates some of the practical challenges of spline models."
bayes  statistics  stan 
yesterday by aapl

« earlier    

related tags

$  2017  a/b  abdsc  abortion  advertising  ai  algorithms  americawhatwentwrong  analysis  analytics  analyze  animals  article  attention  bayes  bigdata  blockchain  book  books  bootstrap  brexit  bruegel.eu  bunnies  business  ceos  charts  cheatsheet  cisgender  clojure  code  collection  color  comics  consider:looking-to-see  consider:representation  cost-effectiveness-analysis  cost-effectiveness  course  courses  cross-validation  cute  d&d  dashboard  data-analysis  data  datagovernance  datamining  dataquality  datascience  dataviz  deeplearning  design  development  domestic_violence  dplyr  economics  education  empirical_processes  energy  engagement  english  environment  errors  europe  examples  export  extinction  facebook  fantasyfootball  feature-construction  finance  fivethirtyeight  food  form  forms  free  fun  gender  gerrymandering  github  graphing  graphs  have_read  health  history  howto  humanity  idlib  idps  immigration  import  inference  inference_to_latent_objects  installer  interactive  iot  iphone  irs  johnroberts  judiciary  kids  language  leaders  learn  learning  lego  lgbtq  library  lingotax  linguistics  literature  lo  long-range_dependence  machine-learning  machinelearning  majestyofthelaw  malinda  mar15  math  mathematics  matplotlib  metrics  migration  modeling  modelling  money  mooc  movies  multilevel  national  nba  network-theory  nlp  nlproc  nltk  nonparametric  nudge-targets  numerical-methods  office  ons  open-source  optimism  oracle  p-values  packages  paper  parser  percentile  performance  plainlanguage  policy  politics  population  precisionjournalism  predictions  predictive  pregnancy  probability  product  programming  pseudoscience  psychology  public_opinion  pydata  python  r-project  r  rails  rather-interesting  reference  religious  reporting  reproductiverights  research  rpgs  rust  sample  sampling  science  sex  significance  socialmedia  spatial-statistics  spatio-temporal-statistics  spectroscopy  sports  stackoverflow  stan  stats  stochastic_processes  stock  supremecourt  survey  survival  time_series  to:nb  trade  trans  transgender  trends  tutorial  twitter  uk  ux  variable-selection  video  virtualmachine  visualization  vm  webtools  wikipedia  wind  with  ya 

Copy this bookmark:



description:


tags: