Questions in topic: "confidence"
https://developer.ibm.com/answers/questions/topics/single/264448.html
The latest questions for the topic "confidence"Add confidence intervals to scatterplot
https://developer.ibm.com/answers/questions/528112/add-confidence-intervals-to-scatterplot.html
Hello everyone. An editor has asked us to add confidence intervals in our scatterplot. I was wondering to select the function 95% CIs mean which draws two lines. Do you think this procedure is correct? Have you other suggestions?confidenceMon, 03 Feb 2020 23:06:14 GMTscandoneHow is calculated the Confidence in IBM Watson Assistant?
https://developer.ibm.com/answers/questions/520657/how-is-calculated-the-confidence-in-ibm-watson-ass.html
I'm trying understand how the IBM Watson assistant calculate the confidence of intent recognition. I already searched in the forum, but I didn't find anything explaining with relation with Deep Learning. I want to know how is calculate it, the Deep Learning Parameters ( weight matrix, activation function...) used in the math formula.watson-assistantconfidencemathTue, 22 Oct 2019 16:42:16 GMTcarlleston“Score”, “Confidance” and "search order"at Watson Discovery
https://developer.ibm.com/answers/questions/516341/score-confidance-and-search-orderat-watson-discove.html
Can you tell me about Watson Discovery's score and Confidance?
Question 1
・ Is it good to recognize that the score is high when the search keyword decomposed into morphemes has a large amount of matches in the search text?
• Or are the ones with a high percentage of matches?
・ Can you tell me otherwise?
Question 2
Is it correct that the search order is determined by the score and relevance training?
If the recognition is correct, what is the contribution to the order of the score and the confidance?
Question 3
・ Does the score change when learning answers? Will there be any changes as a result of annotations in Knowledge Studio?spssdiscoverywatson servicesconfidencescoreMon, 02 Sep 2019 01:29:33 GMTshigematsuHow to calculate bias-corrected bootstrapped confidence interval for PARTIAL SPEARMAN correlations?
https://developer.ibm.com/answers/questions/512902/how-to-calculate-bias-corrected-bootstrapped-confi.html
I used the code below and ran partial Spearman correlations:
NONPAR CORR
/MISSING = LISTWISE
/MATRIX OUT(*).
RECODE rowtype_ ('RHO'='CORR') .
PARTIAL CORR
/significance = twotail
/MISSING = LISTWISE
/MATRIX IN(*).
How can I calculate bias-corrected bootstrapped confidence intervals for this? I tried, but it did not work in conjunction with the /MATRIX IN(*) command. Any advice appreciated.
TIAcorrelationbootstrapconfidencebootstrappingpartialFri, 26 Jul 2019 15:33:24 GMTDGolmhow can I compute confidence for correct and incorrect answers?
https://developer.ibm.com/answers/questions/507149/how-can-i-compute-confidence-for-correct-and-incor.html
Hi,
In my survey each participant had 8 items to judge with a forced choice, after recoding them with 1 for correct and 0 for incorrect and they had to give their confidence out of a scale of 7. I have already computed the total scores for correct answers, but I don't know how to find the confidence for these and for the remaining ones (incorrect).
I hope is clear..thank you for your help :)spsshow-toconfidencescorecomputingWed, 05 Jun 2019 18:50:15 GMTAuroraA.[Watson Assistant] Intent confidence differ in the 'try it' console vs api
https://developer.ibm.com/answers/questions/501949/watson-assistant-intent-confidence-differ-in-the-t.html
When comparing the confidence that Assistant gives when querying a specific message, we are getting distinct values when consuming the service from the API and the 'Try It' console. One of the cases that we found this, the values were 0.0065 vs 0.574, a huge difference. The message and workspace that we used are exactly the same in both. Is this an expected behavior of the service? Are there something that we are missing?
Thanks!assistantconfidenceThu, 18 Apr 2019 17:01:04 GMTCarolina BustosSPSS Statistics Confidence Interval
https://developer.ibm.com/answers/questions/500635/spss-statistics-confidence-interval.html
I created a simple model, two rows and two columns. I need to show confidence intervals. I added lower confidence count and upper confidence count. They appear fine. The accompanying percentages appear fine. But no matter what I do, the percentages do not add up to 100%. How do I make them do that?confidenceWed, 10 Apr 2019 02:58:19 GMTElmoPuddySPSS Output not giving upper CI
https://developer.ibm.com/answers/questions/499887/spss-output-not-giving-upper-ci.html
When completing my multivariate log regression my output looks like this:
Odds Ratio Confidence Interval (95%)
0.000 0.000- .
The n for this variable is very small, is this why? How should I record this in my write up?spssstatisticsconfidencedifferencenegativeboundWed, 03 Apr 2019 09:49:15 GMTMWBARTSUnexpected match in Watson Assistant skill
https://developer.ibm.com/answers/questions/495428/unexpected-match-in-watson-assistant-skill.html
I have created a skill in Watson Assistant with the following intent defined:
{
"intent": "goodbyes",
"examples": [
{
"text": "good bye"
},
{
"text": "cu"
},
{
"text": "later"
},
{
"text": "exit"
},
{
"text": "quit"
},
{
"text": "have a nice day"
},
{
"text": "farewell"
},
{
"text": "bye now"
},
{
"text": "cya"
},
{
"text": "see you"
},
{
"text": "adieu"
},
{
"text": "stop"
},
{
"text": "I'm leaving"
},
{
"text": "we are done"
},
{
"text": "bye"
},
{
"text": "so long"
},
{
"text": "ciao"
},
{
"text": "au revoir"
}
],
"description": ""
},
When I enter the term "again" or "back" to try out the skill, WA returns a match to this intent with a nearly .9 level of confidence. I have fixed this problem by marking those terms as irrelevant. However, I cannot understand *why there is any match at all to the intent*, given that there is no structural or semantic overlap between my training terms and the terms "again" or "back". Nor can I understand *why the confidence level is so high* for what is obviously a mistake for use miserable humans.
I could send the entire json file of the skill if I had a private address to which to send it. I prefer to not post it in this public place.watson-assistantconfidencematchMon, 04 Mar 2019 08:17:25 GMTJosiah CarberryHow to get exakt confidence interval for kategorial data?
https://developer.ibm.com/answers/questions/488408/how-to-get-exakt-confidence-interval-for-kategoria.html
How can i get 95% confidence interval in SPSS for categorial data? I need to get them for frequencies of a categorial data. For example: population xy had 70% positivity in blood cultures.
I know the way for continous data in "descriptive analyse" - "explorative data analysis" but not for categorial data and their frequency.
Look forward for answering and thank youconfidenceTue, 15 Jan 2019 16:52:16 GMTlaurawalter12595% confidence intervals for chi square with 2 IVs?
https://developer.ibm.com/answers/questions/479398/95-confidence-intervals-for-chi-square-with-2-ivs.html
I'm conducting a chi-square analysis with two dichotomous IVs and one dichotomous DV. Is there a way to obtain 95% confidence intervals in SPSS for the IVs at both levels of the other IV?spssconfidenceintervalThu, 08 Nov 2018 01:02:45 GMTKatjaPConfidence gets low after triggering Dialog having sub nodes
https://developer.ibm.com/answers/questions/478799/confidence-gets-low-after-triggering-dialog-having.html
We are using IBM Watson Assistant
If we give back to back inputs that triggers a dialog having sub nodes.
Watson start to give us response with very low confidence.
How to resolve this?watson-assistantconfidencelowMon, 05 Nov 2018 14:49:19 GMTnaveedhanHow to deal with an input having two question in Watson Assistant
https://developer.ibm.com/answers/questions/478794/how-to-deal-with-an-input-having-two-question-in-w.html
We are using IBM Watson Assistant in Asp.Net Application.
If user give an input like following:
**Hi, How are you? Can you please assist me to log in?”**
In this input there are examples from two different **intents**.
Watson picks up the intent **#howareyou** and respond with a confidence of **0.61524133682250981** But we have a policy to show output to user when confidence is equal or greater than 0.7 otherwise ask the question again.
How can we handle this type of case.watsonassistantchatbotinputconfidenceMon, 05 Nov 2018 14:23:23 GMTnaveedhanHow do i get confidence intervals for the Harrell's C index values obtained using spss syntax from http://www-01.ibm.com/support/docview.wss?uid=swg21478383 ?
https://developer.ibm.com/answers/questions/459874/how-do-i-get-confidence-intervals-for-the-harrells.html
How do i get confidence intervals for the Harrell's C index values obtained using spss syntax from http://www-01.ibm.com/support/docview.wss?uid=swg21478383 ?confidenceintervalSun, 22 Jul 2018 00:34:54 GMTHans36How do I create an error bar chart using bootstrapped 95% confidence intervals?
https://developer.ibm.com/answers/questions/455254/how-do-i-create-an-error-bar-chart-using-bootstrap.html
I understand how to create a general error bar chart, but I would like to use specific bootstrap confidence intervals and there doesn't seem to be an option for this. Am I missing something? or is that not an option?spssstatisticsspssstudentspss25bootstrapchartsconfidencebootstrappingMon, 25 Jun 2018 20:41:31 GMTkieraaDatacap XML Layout
https://developer.ibm.com/answers/questions/448397/datacap-xml-layout.html
Hi, I would like to know how the ""cn" value is calculated in Datacap xml layout. Could anyone help me to understand this? In the below example, the recognition seems to be very good however the word confidence is 9599 and the character confidence is 10 and 6. How is it being calculated?
w pos="41,1488,68,1498" cn="9599" s="2" v="vice"
C pos="41,1488,44,1438" cn="10" s="2" v="v"
C pos="51,1495,55,1434" cn="6" s="2" v="i"
C pos="53,1486,62,1423" cn="10" s="2" v="c"
C pos="61,1489,67,1498" cn="10" s="2" v="e"
Also what does "V" indicates?
Many Thanks!watsonwatson-assistantconversationdatacap9.1confidenceThu, 17 May 2018 14:46:46 GMTJujoo_2017Changing the confidence interval in nonlineair regression
https://developer.ibm.com/answers/questions/448026/changing-the-confidence-interval-in-nonlineair-reg.html
Dear Forum,
Currently I am fitting data with a function using "Analyze--> Regressions--> Nonlinear Regression" in SPSS. I use a function with 3 parameters to be estimated. There are no parameter constraints. In "Options" I use the default Levenberg-Marquardt estimation method. This provides me with a function fit with 3 estimated parameters with 95% confidence intervals. I would like to have SPSS compute 99 % or 99,99 % confidence intervals using this method.
However, unlike the tab "Regression-->Linear", there is no tab named "Statistics" at which you can set the confidence interval to be computed. Have you got any ideas how I can change the confidence value of the computed confidence interval in nonlinear regression, using either buttons or syntax code?
Thanks in advance,
LucasregressionconfidenceWed, 16 May 2018 10:01:14 GMTLucasStamWrong confidence for multiple intents
https://developer.ibm.com/answers/questions/441761/wrong-confidence-for-multiple-intents.html
I've got 4 intents:<br>
#buy_dryer (I want to buy dryer)<br>
#buy_vacuum (I want to buy vacuum)<br>
#register_dryer (I want to register dryer)<br>
#register_vacuum (I want to register vacuum)<br>
I don't have any entity.<br>
I've got following dialogs:<br>
welcome (Hello. How can I help you?)<br>
#buy_dryer (You want to buy dryer)<br>
#buy_vacuum (You want to buy vacuum)<br>
#register_dryer (You want to register dryer)<br>
#register_vacuum (You want to register vacuum)<br>
anything_else (I didn't understand. You can try rephrasing.)<br>
When I start chat with "I want to buy dryer. I want to register vacuum" it returns response:
{
"output": {
"text": [
"You want to register vacuum"
],
"nodes_visited": [
"node_4_1523418978260"
],
"log_messages": []
},
"input": {
"text": "I want to buy dryer. I want to register vacuum"
},
"intents": [
{
"intent": "register_vacuum",
"confidence": 0.25246544480323796
},{
"intent": "register_dryer",
"confidence": 0.22784374952316286
},{
"intent": "buy_vacuum",
"confidence": 0.20793357565999032
},{
"intent": "buy_dryer",
"confidence": 0.20648838430643082
}
],
...
My question is why #buy_dryer intent's confidence is too low? I expected #buy_dryer and #register_vacuum will be together on the top. I want to handle multiple intents in one sentence.<br>
Other low priority question is why confidence is too low for all intents. Confidence is 0.2 not 0.7.watsonicnwatson-assistantconversationconfidenceWed, 11 Apr 2018 06:18:35 GMTbob57Can I directly change the confidence of a result?
https://developer.ibm.com/answers/questions/419184/can-i-directly-change-the-confidence-of-a-result.html
Hi,
Can I directly change the confidence of a result?
I found that even though I set some results as 'Not Relevant', the confidence score is still very high. Also, even though I set some results as 'relevant', the confidence is still not very high. Do I have an easier way to directly change the confidence?watson-discoveryconfidencerelevanceFri, 15 Dec 2017 09:00:06 GMTbjct1990Disable low confidence warning during Verify.
https://developer.ibm.com/answers/questions/418477/disable-low-confidence-warning-during-verify.html
I need to use confidence levels in my Datacap application, but i don't like low confidence warning message. It is ok that low confidence fields are yellow, but additional message is frustrating and time wasting. Is any way to disable the message without disabling confidence levels at all?
DC 9.1 with FastDoc and ICN.datacapconfidenceWed, 13 Dec 2017 08:25:17 GMTadiiikHow could I make Watson to trigger a specific intent every time when it gets the confidence level less than 0.6?
https://developer.ibm.com/answers/questions/417256/how-could-i-make-watson-to-trigger-a-specific-inte.html
How could I train Watson to trigger a specific intent (Irrelevant) whenever the user type something that identified by Watson as less than 0.6 confidence value. I know I could do this on the application code, but I need to be on Watson as the Irrelevant intent is a loop of actions that has to be triggered.watsonwatson-assistantconversationconfidenceWed, 06 Dec 2017 18:22:41 GMTSantosh_SNuse of the ciinterval command to calculate joint confidence interval for MANOVA
https://developer.ibm.com/answers/questions/416744/use-of-the-ciinterval-command-to-calculate-joint-c.html
Hi,
I am having trouble figuring out the proper syntax to input to calculate a joint confidence interval for a MANOVA. So far the syntax is stated below. On other help pages it indicates that there is a /ciinterval command where I can specify "joint(.95)" Where would I input these commands within the syntax noted below.
DATASET ACTIVATE DataSet10.
LOGISTIC REGRESSION VARIABLES ETOHGRP_ORIG
/METHOD=BSTEP(LR) Speed_T AttWmExec_T Learn_T Mem_T Verbal_T Motor_T
/CRITERIA=PIN(.05) POUT(.10) ITERATE(20) CUT(.5)confidenceintervaljointMon, 04 Dec 2017 19:07:40 GMTvebryantConfidence intervals contain 0, but the p-value is significant
https://developer.ibm.com/answers/questions/402840/confidence-intervals-contain-0-but-the-p-value-is.html
In a multilevel mixed effects model, bootstrapped BCa CI's contain zero [-.0013, .04162] but the p < .05?
Is this because they are asymmetrical, or could their be an alternative cause?
SYNTAX
BOOTSTRAP
/SAMPLING METHOD=STRATIFIED(STRATA=Forum )
/VARIABLES TARGET=AbsolutistLog10 INPUT=Condition Forum WC
/CRITERIA CILEVEL=95 CITYPE=BCA NSAMPLES=1000
/MISSING USERMISSING=EXCLUDE.
MIXED AbsolutistLog10 BY Forum Condition
/CRITERIA=CIN(95) MXITER(150) MXSTEP(10) SCORING(1) SINGULAR(0.000000000001) HCONVERGE(0,
ABSOLUTE) LCONVERGE(0, ABSOLUTE) PCONVERGE(0.000001, ABSOLUTE)
/FIXED=Condition | SSTYPE(3)
/METHOD=REML
/REGWGT=WC
/PRINT= SOLUTION TESTCOV
/RANDOM=Forum(Condition) | SUBJECT(Forum) COVTYPE(VC).confidenceintervalp-valueSat, 23 Sep 2017 19:36:27 GMTMoe1986How do I crop out a snippet from a document based on the coordinates provided by Datacap?
https://developer.ibm.com/answers/questions/401176/how-do-i-crop-out-a-snippet-from-a-document-based.html
Hi,
I would like to ask if it is possible to crop out snippets from documents on Datacap.
When I extract a name - "John Doe" from a document, Datacap will provide the coordinates of where "John Doe" is extracted from a document. I would like to crop out the snippet and store it for future retrieval.
On a side note, where is the best place to store this snippet? E.g. Database, FileNet, file systems
Best regards,
Aaron Chiadatacapfilenetdatabasenavigatorimagesdatacap9.0.1confidenceFri, 15 Sep 2017 08:53:43 GMTaaronchiaHow to send DataCap batch in FixUp Job when confidence level is less than 5?
https://developer.ibm.com/answers/questions/399622/how-to-send-datacap-batch-in-fixup-job-when-confid.html
I want to check Confidence level of Document and if it is less than 5 send it to FixUp job?datacapdatacap9.0.1confidencebatch jobdatacapstudiofixupFri, 08 Sep 2017 13:39:58 GMTPujaJagtapSPSS one sample t test displays incorrect CI bounds for diff of means
https://developer.ibm.com/answers/questions/387843/spss-one-sample-t-test-displays-incorrect-ci-bound.html
I ran a simple one sample t-test. All of the results are correct, except for the 95% Confidence Interval of the difference. The lower and upper bounds are both negative. It seems to me that each of the bounds should have the same magnitude, but with different signs; so the lower bound should be negative and the upper bound should be positive. For sure, 0, indicating no difference between the means should reside inside of this confidence interval.
N = 10; Mean = 77.7; Std Dev = 11.46056; St Error = 3.62415
Here is the output:
One-Sample Test
Test Value = 89
t = -3.118; df = 9; Sig. (2-tailed) = .012
; Mean Difference = -11.30000
95% Confidence Interval of the Diff : Lower = -19.4984 ;
Upper = -3.1016
Please let me know what the numbers represent in the output because it does not represent the 95% Confidence Interval of the difference of the means. Thank you!spssstatisticsconfidencedifferencenegativeboundWed, 19 Jul 2017 21:31:27 GMTjsbaucom