Talk:Most frequent k chars distance: Difference between revisions

← Older edit

Talk:Most frequent k chars distance (view source)

Revision as of 15:21, 30 May 2014

439 bytes added , 9 years ago

→‎Move page?: seconded

Anonymous user

rosettacode>DanBron

Revision as of 22:46, 29 May 2014 (view source) rosettacode>DanBron (sign comments) ← Older edit		Latest revision as of 15:21, 30 May 2014 (view source) rosettacode>DanBron (→‎Move page?: seconded)
(5 intermediate revisions by 2 users not shown)
Line 1: == Move page? == I'm thinking that the page should not be a draft task/task and moved to this sub-page of the authors: "Shedai/Most frequent k chars distance" until it is ready to become a draft task. I would like to do this in the next three days unless there are reasoned objections. --[[User:Paddy3118\|Paddy3118]] ([[User talk:Paddy3118\|talk]]) 07:40, 30 May 2014 (UTC) :Seconded --[[User:DanBron\|DanBron]] ([[User talk:DanBron\|talk]]) 15:21, 30 May 2014 (UTC) == What is this? == Line 26 ⟶ 33: ::::6. The alternative interpretation (taken by both current implementations) is that the frequencies ''are'' actually summed, and this makes <tt>MostFreqKSDF(str1,str2,2,100) = 83</tt> work, but that breaks <tt>MostFreqKSDF('night','nacht',2,10)=9</tt> (which would be <tt>8</tt>), <tt>MostFreqKSDF('research', 'research',2,10)=8</tt> (would be <tt>6</tt>), <tt>MostFreqKSDF('aaaaabbbb','ababababa',2,10)=1</tt> (would be <tt>-8</tt>), <tt>MostFreqKSDF('significant', 'capabilities', 2, 10)=5</tt> (would be <tt>4</tt>), and the original example, given (implicitly) in the preamble, <tt>MostFreqKSimilarity('r2e2','e2s1',2,10)=8</tt> (would be <tt>6</tt>). ::::7. Just to make sure I wasn't crazy, I ran the Java code posted by the original author of the task, and as expected (~~but~~ for the reasons outlined just above) its outputs disagree with the values given in the test cases (for example, it gives <tt>-8</tt> for ~~the~~ <tt>'aaaaabbbb'</tt> vs <tt>'ababababa'</tt> ~~example~~, and actually crashes for ~~the~~ <tt>'my'</tt> vs <tt>'a'</tt> ~~example~~). ::: Anyway, unless I'm overlooking something obvious, an implementation can not be consistent with both the pseudocode and the required examples. Anyone care to point me in the right direction? ::: (For now, I'm going to post the J code I wrote based on the "frequencies are compared" interpretation, which, IMO, accords better with both the test cases and the ~~actualy~~actual concept of "similarity".) ::: --[[User:DanBron\|DanBron]] ([[User talk:DanBron\|talk]]) 22:46, 29 May 2014 (UTC)