Scribd
Upload
346 views

Bayesian Statistics For Data Science - Towards Data Science

This document summarizes key concepts in Bayesian statistics and their significance for data science. It compares Bayesian and frequentist approaches, explaining Bayesian inference and the use of priors, likelihoods, and posteriors. The document also discusses Bayesian tests for significance like Bayes factors and High Density Intervals, which are preferable to p-values and confidence intervals. Overall, the document argues that Bayesian statistics encompasses models useful for data science, especially when data is limited or strong prior intuitions exist.

Uploaded by

Dom DeSicilia
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
346 views

Bayesian Statistics For Data Science - Towards Data Science

This document summarizes key concepts in Bayesian statistics and their significance for data science. It compares Bayesian and frequentist approaches, explaining Bayesian inference and the use of priors, likelihoods, and posteriors. The document also discusses Bayesian tests for significance like Bayes factors and High Density Intervals, which are preferable to p-values and confidence intervals. Overall, the document argues that Bayesian statistics encompasses models useful for data science, especially when data is limited or strong prior intuitions exist.

Uploaded by

Dom DeSicilia
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Ankit Rathi Follow

Data Science Architect | Kaggle Expert | Data Blogger | All views personal |
https://about.me/ankitrathi
Aug 15 · 5 min read

Bayesian Statistics for Data Science

Bayesian Statistics for Data Science

This is the 5th post of blog post series ‘Probability & Statistics for Data
Science’, this post covers these topics related to Bayesian statistics and their
signi cance in data science.

• Frequentist Vs Bayesian Statistics

• Bayesian Inference

• Test for Signi cance

• Signi cance in Data Science

. . .

Frequentist Vs Bayesian Statistics

Frequentist Statistics tests whether an event (hypothesis) occurs or not.


It calculates the probability of an event in the long run of the
experiment. A very common aw found in frequentist approach i.e.
dependence of the result of an experiment on the number of times the
experiment is repeated.

Frequentist statistics su ered some great aws in its design and


interpretation which posed a serious concern in all real life problems:

1. p-value & Con dence Interval (C.I) depend heavily on the sample
size.

2. Con dence Intervals (C.I) are not probability distributions

Bayesian statistics is a mathematical procedure that applies


probabilities to statistical problems. It provides people the tools to
update their beliefs in the evidence of new data.

Frequentist Vs Bayesian Statistics

Frequentist vs. Bayesian


Inference - The Basics of
Bayesian Statistics | Coursera
This course describes Bayesian
statistics, in which one's inferences
about parameters or hypotheses are
updated as…
www.coursera.org

Bayesian Inference

To understand Bayesian Inference, you need to understand Conditional


Probability & Bayes Theorem, if you want to review these concepts,
please refer my earlier post in this series.

Probability for Data Science


This is the 2nd post of blog post
series ‘Probability & Statistics for
Data Science’, this post covers these
topics…
towardsdatascience.com

Bayesian inference is a method of statistical inference in which Bayes’


theorem is used to update the probability for a hypothesis as more
evidence or information becomes available.

An important part of Bayesian Inference is the establishment of


parameters and models. Models are the mathematical formulation of
the observed events. Parameters are the factors in the models a ecting
the observed data. To de ne our model correctly , we need two
mathematical models before hand. One to represent the likelihood
function and the other for representing the distribution of prior beliefs .
The product of these two gives the posterior belief distribution.
Courtesy: http://jason-doll.com/wordpress/?page_id=127

How Bayes Theorem works

Likelihood Function

A likelihood function is a function of the parameters of a statistical


model, given speci c observed data. Probability describes the
plausibility of a random outcome, without reference to any observed
data while Likelihood describes the plausibility of a model parameter
value, given speci c observed data.

Likelihood function

Likelihood function - Wikipedia


In Bayesian inference, although one
can speak about the likelihood of
any proposition or random variable
given another…
en.wikipedia.org

Prior & Posterior Belief distribution

Prior Belief distribution is used to represent our strengths on beliefs


about the parameters based on the previous experience. Posterior Belief
distribution is derived from multiplication of likelihood function & Prior
Belief distribution.

As we collect more data, our posterior belief move towards prior belief
from likelihood:

Courtesy: https://jimgrange.wordpress.com/2016/01/18/pesky-priors/

(Pesky?) Priors
When I tell people I am learning
Bayesian statistics, I tend to get one
of two responses: either people look
at me…
jimgrange.wordpress.com

Test for Signi cance

Bayes factor

Bayes factor is the equivalent of p-value in the Bayesian framework. The


null hypothesis in Bayesian framework assumes ∞ probability
distribution only at a particular value of a parameter (say θ=0.5) and a
zero probability else where. The alternative hypothesis is that all values
of θ are possible, hence a at curve representing the distribution.

Courtesy: http://areshenk-research-notes.com/bayes-factors-and-stopping-rules/

Using Bayes Factor instead of p-values is more bene cial in many cases
since they are independent of intentions and sample size.

Replacing p-values with Bayes-


Factors: A Miracle Cure for the
Replicability Crisis in
Psychological…
How Science Should Work Lay
people, undergraduate students, and
textbook authors have a simple
model of science…
replicationindex.wordpress.com
High Density Interval (HDI)

High Density Interval (HDI) or Credibility Interval is equivalent to


Con dence Interval (CI) in Bayesian framework. HDI is formed from the
posterior distribution after observing the new data.

Courtesy: https://www.slideshare.net/ASQwebinars/bayesian-methods-in-reliability-engineering-
15204318

Using High Density Interval (HDI) instead of Con dence Interval (CI) is
more bene cial since they are independent of intentions and sample
size.

Con dence vs. Credibility


Intervals
Tomorrow, for the nal lecture of the
Mathematical Statistics course, I will
try to illustrate - using Monte
Carlo…
freakonometrics.hypotheses.org

Moreover, there is a nice article published on AnalyticsVidhya on this


which elaborate on these concepts with examples:

Bayesian Statistics explained to


Beginners in Simple English
Introduction Bayesian Statistics
continues to remain
incomprehensible in the ignited
minds of many analysts. Being…

www.analyticsvidhya.com

Signi cance in Data Science

Bayesian statistics encompasses a speci c class of models that could be


used for Data Science. Typically, one draws on Bayesian models for one
or more of a variety of reasons, such as:

• having relatively few data points

• having strong prior intuitions

• having high levels of uncertainty

And there are scenarios where Bayesian statistics will perform


drastically, please read following discussion for details:
What's the relationship between
bayesian statistics and machine
learning?
Answer (1 of 2): Machine learning is a
broad eld that uses statistical
models and algorithms to
automatically learn…

www.quora.com

References:

Bayesian Methods for Hackers


The Bayesian method is the natural
approach to inference, yet it is
hidden from readers behind chapters
of slow…

nbviewer.jupyter.org

Bayesian Statistics: From


Concept to Data Analysis |
Coursera
About this course: This course
introduces the Bayesian approach to
statistics, starting with the concept
of probability…
www.coursera.org

. . .

Thank you for reading my post. If you enjoyed it, please clap button 👏 so
others might stumble upon it. I regularly write about Data & Technology
on LinkedIn & Medium. If you would like to read my future posts then
simply ‘Connect’ or ‘Follow’. Also feel free to listen to me on SoundCloud.

You might also like