User contributions for Sean a wallis
A user with 81 edits. Account created on 17 September 2008.
11 September 2022
- 10:1810:18, 11 September 2022 diff hist +48 Mutual information →For discrete data: The original statement was simply imprecise and liable to mislead. The revised equivalence is true iff the log base is the same. Bear in mind that log-likelihood is quoted in nats (ln), whereas discrete MI is often quoted normalised to a base k.
28 January 2021
- 18:1518:15, 28 January 2021 diff hist +6 User:Sean a wallis No edit summary
- 18:1318:13, 28 January 2021 diff hist +673 N User:Sean a wallis ←Created page with 'Sean Wallis is a Principal Research Fellow at the Survey of English Usage at University College London. He is a researcher in Corpus Linguistics who...'
- 17:5717:57, 28 January 2021 diff hist +10 User:Sean a wallis/sandbox →The Design of Englicious current
- 17:5617:56, 28 January 2021 diff hist +294 User:Sean a wallis/sandbox No edit summary
- 17:4817:48, 28 January 2021 diff hist −78 User:Sean a wallis/sandbox No edit summary
- 17:3017:30, 28 January 2021 diff hist +572 User:Sean a wallis/sandbox No edit summary
27 January 2021
- 20:3620:36, 27 January 2021 diff hist +668 User:Sean a wallis/sandbox No edit summary
- 17:2617:26, 27 January 2021 diff hist +673 User:Sean a wallis/sandbox No edit summary
- 17:1717:17, 27 January 2021 diff hist +24 User:Sean a wallis/sandbox →The 'Grammar Gap'
- 17:0617:06, 27 January 2021 diff hist +9 User:Sean a wallis/sandbox →The 'Grammar Gap'
- 17:0417:04, 27 January 2021 diff hist +319 User:Sean a wallis/sandbox →The 'Grammar Gap'
26 January 2021
- 12:4812:48, 26 January 2021 diff hist +210 Binomial proportion confidence interval →Wilson score interval: Addition of image to show the properties of the Wilson score interval exactly as with the normal approximation interval.
- 12:4312:43, 26 January 2021 diff hist +8 m File:Normal approx interval and logistic example.png No edit summary
- 12:4212:42, 26 January 2021 diff hist +750 N File:Wilson score interval and logistic example.png == Summary == Wallis, Sean. Original image by author. A version of this plot is published in Wallis (2021: 180).<ref name=Wallis2021> {{cite book | last = Wallis | first = Sean A. | title = Statistics in Corpus Linguistics - a new approach | publisher = Routledge | location = New York | isbn = 9781138589384 | url = https://www.routledge.com/Statistics-in-Corpus-Linguistics-Research-A-New-Approach/Wallis/p/book/9781138589384 | date = 2021 }}</ref> Wilson score intervals for ''n'' = 10...
- 12:2412:24, 26 January 2021 diff hist +204 Binomial proportion confidence interval →Normal approximation interval: Addition of image plotting the interval along the logistic curve. Anyone who wishes to employ this interval needs to understand its properties, and this is a simple data-neutral visualisation.
- 12:1712:17, 26 January 2021 diff hist +835 N File:Normal approx interval and logistic example.png Wallis, Sean. Original image by author. A version of this plot is published in Wallis (2021: 180).<ref name=Wallis2021> {{cite book | last = Wallis | first = Sean A. | title = Statistics in Corpus Linguistics - a new approach | publisher = Routledge | location = New York | isbn = 9781138589384 | url = https://www.routledge.com/Statistics-in-Corpus-Linguistics-Research-A-New-Approach/Wallis/p/book/9781138589384 | date = 2021 }}</ref> Gaussian (normal approximation) intervals for n = 1...
25 January 2021
- 18:0518:05, 25 January 2021 diff hist +47 m Binomial proportion confidence interval →Wilson score interval
- 18:0218:02, 25 January 2021 diff hist +333 Binomial proportion confidence interval →Wilson score interval: Added to the preliminary to motivate the interval.
- 17:4917:49, 25 January 2021 diff hist +34 m Binomial proportion confidence interval →Wilson score interval with continuity correction: clarifying that the average probability is the average *coverage* probability, and that the 'nominal value' is 1-alpha, i.e. a 95% CI has a coverage prob = 0.95.
- 17:4517:45, 25 January 2021 diff hist +120 m Binomial proportion confidence interval →Wilson score interval: URL link to publisher's website
24 January 2021
- 12:1112:11, 24 January 2021 diff hist +138 Binomial proportion confidence interval →Wilson score interval
- 11:5911:59, 24 January 2021 diff hist +16 m Binomial proportion confidence interval →The interval equality principle: using lower case 'normal' for consistency with Wikipedia style
- 11:5511:55, 24 January 2021 diff hist −6 m Binomial proportion confidence interval →Comparison of different intervals
- 11:5511:55, 24 January 2021 diff hist +20 m Binomial proportion confidence interval →Comparison of different intervals
- 10:4410:44, 24 January 2021 diff hist +412 Binomial proportion confidence interval →Comparison of different intervals
- 10:3210:32, 24 January 2021 diff hist +3 m Binomial proportion confidence interval →Wilson score interval
- 10:2510:25, 24 January 2021 diff hist +16 m Binomial proportion confidence interval →Wilson score interval: introducing explicit interval terms (w-, w+) into equation for Wilson solution (far easier for reader than leaving implicit)
- 10:1610:16, 24 January 2021 diff hist +1,902 Binomial proportion confidence interval →Wilson score interval: Added - section on interval equality principle (extremely useful in practice) with image + more explanation of continuity-corrected interval and simpler derivation.
- 08:5208:52, 24 January 2021 diff hist −1 m File:Wilson score pdf and interval equality.png →Summary current
- 08:5108:51, 24 January 2021 diff hist +617 N File:Wilson score pdf and interval equality.png Wallis, Sean. Original image by author, method in https://corplingstats.wordpress.com/2018/09/25/plotting-the-wilson-distribution, also published in Wallis (2021: 297-313). Example probability density function (pdf) for Wilson score distribution, and Normal pdfs at interval bounds (observed proportion = 0.3, sample size n = 10, two-tailed error level alpha = 0.05). Demonstrates the equality of tail areas (highlighted), termed the 'interval equality principle' (Wallis 2021). <ref>Wallis, S.A...
5 January 2021
- 11:5511:55, 5 January 2021 diff hist +5,363 N User:Sean a wallis/sandbox ←Created page with ''''Englicious''' is a website compendium of resources, developed by a team of linguists at the Survey of English Usage at University College London, to...'
26 March 2019
- 14:2114:21, 26 March 2019 diff hist −4 m Chi-squared test →See also
- 14:2014:20, 26 March 2019 diff hist +11 m Chi-squared test →See also
- 14:1614:16, 26 March 2019 diff hist +966 Chi-squared test →See also: related the 2 x 1 and 2 x 2 tests to single-interval and dual-interval (difference) tests
- 13:4813:48, 26 March 2019 diff hist +126 Chi-squared test →Other examples of chi-squared tests: Reorganised, and added binomial test
- 13:4413:44, 26 March 2019 diff hist +23 Chi-squared test →Fisher's exact test: specify the purpose for which the test is used
- 13:4313:43, 26 March 2019 diff hist −2 m Chi-squared test →Other examples of chi-squared tests: Moved to higher level - these are not goodness of fit tests (e.g. the Binomial test is a gof test, but Fisher is not.)
- 13:4213:42, 26 March 2019 diff hist +6 m Chi-squared test →Pearson's chi-squared test: highlight goodness of fit vs. independence and homogeneity - these are different tests, yet students get this wrong all the time, especially in citation!
- 13:3713:37, 26 March 2019 diff hist +146 Chi-squared test →See also: added reference to Wilson score interval
- 13:2513:25, 26 March 2019 diff hist +314 Binomial proportion confidence interval →Agresti–Coull interval: Specify the relationship between Agresti-Coull and Wilson/Normal
- 13:0513:05, 26 March 2019 diff hist +222 Binomial proportion confidence interval →Comparison of different intervals: The original text was rather inconclusive. It makes sense to direct readers to the conclusions of papers.
12 July 2013
- 17:2117:21, 12 July 2013 diff hist +54 Binomial proportion confidence interval →Normal approximation interval
- 17:1717:17, 12 July 2013 diff hist +143 Continuity correction →See also: Add Wilson cc ref
- 17:1617:16, 12 July 2013 diff hist +144 m Yates's correction for continuity →See also: Add Wilson cc ref
- 17:1217:12, 12 July 2013 diff hist +2 m Binomial proportion confidence interval →Wilson score interval with continuity correction: indented
- 14:3614:36, 12 July 2013 diff hist +365 Talk:Binomial proportion confidence interval →suggestions for improvement
- 14:3214:32, 12 July 2013 diff hist +1,103 Binomial proportion confidence interval →Wilson score interval: Continuity correction section added
- 09:5409:54, 12 July 2013 diff hist −86 Binomial proportion confidence interval →Normal approximation interval: Minor changes, except that the advice was too generous towards the normal approximation. Modern analyses following Newcombe 1998 argue it is provably incorrect and should be discontinued in favour of the Wilson.
