RC1: Another example where Fuzzy needs refining
The ESV results appear after clicking more > but there is something wrong with the ranking.
Dave
===
Windows 11 & Android 13
Comments
-
This is working as expected. Ranking does not take account for your personal preferred Bible, but is a global ranking across all Bibles with the results displayed in your preferred Bible. The two verses in Psalms shown above have exact matches for that phase in other translations.
Andrew Batishko | Logos software developer
0 -
And FWIW, if you're trying to use Fuzzy Search to find an exact phrase, you're probably using the wrong tool for the job. The intent of Fuzzy Bible Search is to find that verse that you kind of remember but don't recall the exact wording, or maybe you remember it half from the KJV and half from the NIV. (I'm not very familiar with its implementation, but I wouldn't be surprised if it simply ignores any quotes in your input, because finding exact phrases isn't its job.)
0 -
Perhaps Php 2:12 could be ranked a bit higher in a Fuzzy Bible Search for:
God helps those who help themselves
Keep Smiling [:)]
0 -
I wasn't after an exact phrase, but it's obvious that the fuzzy words i used (not a deliberate test set) came from my preferred bible. So I think ranking should consider preferred bible more highly, and also consider my other prioritised bibles. In any case Fuzzy should list the sources for each of the top five passages.
Dave
===Windows 11 & Android 13
0 -
The "source" is not a single actual Bible. If every Bible in our library except for your preferred Bible included an exact match of your phrase, it's likely to get ranked very high, and wouldn't have a single "source". Likewise, we factor knowledge of the places where Biblical entities are referenced to influence the rank. This is not a "source" that is specific to any one Bible, and in fact may lead to a situation where none of the Bibles have the exact phrase, but a verse is highly ranked because of a high correlation in entities.
There has been some internal discussion about the feasibility of influencing ranking according to your preferred Bible, so the possibility does exist for that to be implemented at some point in the future. I don't have any information as to when or if this will happen.
Andrew Batishko | Logos software developer
0 -
For myself, the translation that I most likely used when I memorized the passage not my current preferred Bible is the one I'd want given priority. I'd be more comfortable with my top 5 getting a ranking boost rather than my single preferred Bible.
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
The "source" is not a single actual Bible.
I requested the "sources for each of the....passages".
There has been some internal discussion about the feasibility of influencing ranking according to your preferred Bible
When I use the words without quotes, the Fuzzy results are unchanged but the ESV results include 1 Sam 25:26 & Prov 14:22, 28:5 which are not amongst the expanded Fuzzy results.
Dave
===Windows 11 & Android 13
0 -
MJ. Smith said:
... my top 5 getting a ranking boost rather than my single preferred Bible
+1 [Y] for Top 5 Bibles receiving a bit of ranking boost.
Dave Hooton said:The "source" is not a single actual Bible.
I requested the "sources for each of the....passages".
Integration with Ref Tagger would be awesome so hovering mouse over Fuzzy Bible Reference has a pop-up with Bible source, which can avoid not licensed message). Dreaming of reference click opening Bible source when my Library has license (and resource is not hidden) or highest prioritized Bible with verse when my Library does not have license (or resource is hidden).
Keep Smiling [:)]
0 -
Dave Hooton said:
The "source" is not a single actual Bible.
I requested the "sources for each of the....passages".
In your example screenshot, even the ESV is a source for the first 5 results. In fact, probably all of the Bibles are a source of some amount of the ranking. Perhaps you are wanting to know which Bibles include that exact phrase, but as Bradley mentioned above, if you are looking for that information, you probably should be doing a Bible search of that phrase on all Bibles. If this is not what you are trying to get at, then I'll need you to give me a little more detail as to what you are looking for.
Dave Hooton said:There has been some internal discussion about the feasibility of influencing ranking according to your preferred Bible
When I use the words without quotes, the Fuzzy results are unchanged but the ESV results include 1 Sam 25:26 & Prov 14:22, 28:5 which are not amongst the expanded Fuzzy results.
I'll pass this along to the right people.
Andrew Batishko | Logos software developer
0 -
And FWIW, if you're trying to use Fuzzy Search to find an exact phrase, you're probably using the wrong tool for the job.
I wonder if this is a design pitfall, where the developers understand how the program is meant to work, but the users will be left with a less satisfying experience?
Users aren't going to approach things the same way the developer will, and expecting or requiring users to use search tool "A" for this type of search, and search tool "B" for that type of search, only puts the complexity and burden on the end-user, who is left wondering why a) their results aren't accurate/reliable, or b) the verse they want to find isn't showing up.
Making the software easier to use improves customer satisfaction, reduces forum questions, and eliminates any "you're doing it wrong" explanations.
If any sophistication could be kept on the back end, instead of requiring the user to master subtle intricacies of search tools, then we might be able to move towards a simpler search approach where it just works, and more people can "naively" benefit from searching, without requiring a particular level of search term expertise.
Thanks to FL for including Carta and a Hebrew audio bible in Logos 9!
0 -
Dave Hooton said:
I requested the "sources for each of the....passages".
Perhaps you are wanting to know which Bibles include that exact phrase
No! I'm wanting to know where a result comes from when it is poorly represented by the passage in my preferred bible; when my significant words are not highlighted.
e.g. Ps 28:3 "workers of evil"
Ps 26:5 "evildoers"
Ps 14:4 "evildoers"
Ps 125:5 "evildoers"
Pr 10:29 "evildoers" (no highlighting in ESV)
seem to be significant for "who do evil" but are not highlighted.
But I prefer results with most of the exact words to be ranked higher.
Dave
===Windows 11 & Android 13
0 -
Dave Hooton said:Dave Hooton said:
I requested the "sources for each of the....passages".
Perhaps you are wanting to know which Bibles include that exact phrase
No! I'm wanting to know where a result comes from when it is poorly represented by the passage in my preferred bible; when my significant words are not highlighted.
e.g. Ps 28:3 "workers of evil"
Ps 26:5 "evildoers"
Ps 14:4 "evildoers"
Ps 125:5 "evildoers"
Pr 10:29 "evildoers" (no highlighting in ESV)
seem to be significant for "who do evil" but are not highlighted.
But I prefer results with most of the exact words to be ranked higher.
All that highlighting does is show where there are words in your preferred translation that exactly match the search you typed in. They don't always strongly indicate why a passage was or was not included as a high ranking result. That's why there is still some internal discussion over whether or not to include this highlighting.
There unfortunately is not a way to indicate where a result comes from, because the result comes from an aggregate of all the Bibles in the system plus additional other information. If you want this information, your best bet is to search all Bibles without the quotes: those who do evil
Andrew Batishko | Logos software developer
0 -
May I ask a question that may related to how this feature will be received? How do the results provided by this algorithm compare to the results provided in L3?
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
MJ. Smith said:
May I ask a question that may related to how this feature will be received? How do the results provided by this algorithm compare to the results provided in L3?
See Eli's response here: https://community.logos.com/forums/p/132639/862877.aspx#862877
Basically, this is doing something completely different than L3's fuzzy search.
Andrew Batishko | Logos software developer
0 -
Basically, this is doing something completely different than L3's fuzzy search.
If the difference in how they work means that the results are radically different then this requires very special treatment in the announcement or the people who have been pushing for the return of the fuzzy search will be disappointed/upset. If the results are similar although the method and constraints are not, this is not a concern.
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
If you want this information, your best bet is to search all Bibles without the quotes: those who do evil
Sadly, this discussion is going round in circles. "If" I wanted that I would have done so. However, there has been some recognition of the issues:-
- consider results in preferred bible(s) as ranking higher
- highlighting is misleading
But there has not been a response to listing the bibles that contribute to the "aggregated" results.
Having used Logos 3's Fuzzy search, I can assure you that similar questions/doubts will arise if this Fuzzy is released as is.
Dave
===Windows 11 & Android 13
0 -
Dave Hooton said:
But there has not been a response to listing the bibles that contribute to the "aggregated" results.
The response is that almost all Bibles contribute to every aggregated result. Listing almost every Bible for almost every result does not seem particularly useful.
Andrew Batishko | Logos software developer
0 -
Dave Hooton said:
But there has not been a response to listing the bibles that contribute to the "aggregated" results.
The response is that almost all Bibles contribute to every aggregated result. Listing almost every Bible for almost every result does not seem particularly useful.
Could you show the versions with, say, greater than .9 confidence? (I don't know what the actual threshold should be, but it seems better than using the highest three confidence Bibles, since you might have several exact matches for a particular phrase.)
Using Logos as a pastor, seminary professor, and Tyndale author
0 -
Justin Gatlin said:Dave Hooton said:
But there has not been a response to listing the bibles that contribute to the "aggregated" results.
The response is that almost all Bibles contribute to every aggregated result. Listing almost every Bible for almost every result does not seem particularly useful.
Could you show the versions with, say, greater than .9 confidence? (I don't know what the actual threshold should be, but it seems better than using the highest three confidence Bibles, since you might have several exact matches for a particular phrase.)
The system does not work in a way that makes this feasible.
Andrew Batishko | Logos software developer
0 -
The system does not work in a way that makes this feasible.
Its ways are quite mysterious.
all sinned and gone astray provides resonable results as it shows that I was mixing two passages Ro 3:23 and Is 53:6. The inclusion of Ro 3:12 is commendable. But sinned and gone astray provides only two results - assuming "we have sinned" whilst ignoring "all have sinned".
Dave
===Windows 11 & Android 13
0
