Search for tag: "inf2-fds"

FDS-S2-05-1-3 Data and code management

We discuss the differing management requirements of data and code.

From  David Sterratt 0 likes 112 plays 0  

FDS-S2-05-1-2 Notebooks versus programs

We will think about the pros and cons of notebooks versus programs

From  David Sterratt 0 likes 119 plays 0  

FDS-S2-05-1-1 Reproducible research

We discuss reproducible research, the motivation for aspects of software engineering practices in data science.

From  David Sterratt 0 likes 141 plays 0  

FDS-S2-03-1-4 Issues in A/B testing

We discuss the issue of statistical versus practical significance, think a little about the ethics of A/B testing, and situate A/B testing in the wider class of statistical inference from 2 samples.

From  David Sterratt 0 likes 153 plays 0  

FDS-S2-03-1-3 Large sample theory of A/B testing

We work through how to apply large-sample theory to A/B testing.

From  David Sterratt 0 likes 171 plays 0  

FDS-S2-03-1-2 Increasing certainty in A/B testing

We look at how to obtain more precise and estimates in A/B testing - and how to avoid some pitfalls.

From  David Sterratt 0 likes 166 plays 0  

FDS-S2-03-1-1 principle of A/B testing

We explain the principle of A/B testing, and how to derive confidence intervals for the difference in population proportions using the bootstrap.

From  David Sterratt 0 likes 202 plays 0  

FDS-S2-02-2-5 Maximum likelihood estimation of logistic regression coefficients

We introduce the principle of maximum likelihood, and show how to derive the likelihood as a function of the coefficients for logistic regression. We also mention one topical application of logistic…

From  David Sterratt 0 likes 161 plays 0  

FDS-S2-02-2-4 The logistic regression classifier

We show how to make a logistic regression classifier, and, in the context of ethics, consider how transparent logistic regression can be made to people whose lives it affects. We also compare…

From  David Sterratt 0 likes 163 plays 0  

FDS-S2-02-2-3 Multiple logistic regression and confidence and intervals

We extend logistic regression to multiple independent variables, and show how we can use the boostrap to estimate confidence intervals for coefficients, and to test hypotheses.

From  David Sterratt 0 likes 181 plays 0  

FDS-S2-02-2-2 Interpretation of logistic regression coefficients

We look at the meaning of logistic regression coefficients, relating them to odds, log odds and odds ratios.

From  David Sterratt 0 likes 197 plays 0  

FDS-S2-02-2-1 Principle of logistic regression

We introduce the principle of logisitc regression, illustrating its application to a credit approval dataset. We also introduce the concepts of odds and odds ratios.

From  David Sterratt 0 likes 229 plays 0  

FDS-S2-02-4 Issues in hypothesis testing

We consider 3 issues in hypothesis testing: Type I and Type II errors; cherry-picking, p-value hacking etc., and a short discussion about the possibility that hypothesis-driven research can hinder…

From  David Sterratt 0 likes 157 plays 0  

FDS-S2-02-1-3 Goodness-of-fit

We'll extend the previous example to a one-way contingency table with multiple categories, in which the chi-squared "goodness-of-fit" statistic is used. We then extend it further to a…

From  David Sterratt 0 likes 223 plays 0  

FDS-S2-02-1-2 P-values

We introduce P-values, an important but tricky concept in hypothesis testing

From  David Sterratt 0 likes 221 plays 0  

FDS-S2-02-1-1 Principle of hypothesis testing

We introduce the principle of hypothesis testing

From  David Sterratt 0 likes 261 plays 0