user avatar
user avatar
user avatar
Jeffrey Girard
  • Member for 6 years, 2 months
  • Last seen this week
109 votes
4 answers
201k views

How do you calculate precision and recall for multiclass classification using confusion matrix?

  • 1,201
46 votes
2 answers
60k views

Determining sample size necessary for bootstrap method / Proposed Method

  • 479
36 votes
3 answers
40k views

Why is AUC higher for a classifier that is less accurate than for one that is more accurate?

  • 1,318
16 votes
4 answers
7k views

Checking whether accuracy improvement is significant

  • 261
10 votes
1 answer
2k views

When is a proper scoring rule a better estimate of generalization in a classification setting?

  • 3,558
6 votes
1 answer
12k views

Can ICC values be negative?

  • 527
6 votes
1 answer
1k views

Fisher's exact test vs kappa analysis

4 votes
1 answer
1k views

How to test whether ICCs from different groups are significantly different?

  • 101
3 votes
0 answers
260 views

How do I calculate reliability between an algorithm and multiple human experts?

  • 31
3 votes
3 answers
2k views

Given a draw from one of two overlapping normal distributions, what is the probability it came from one vs. the other?

2 votes
0 answers
125 views

Assessing the ability of two quantitative methods measuring the same thing

  • 71
1 vote
1 answer
92 views

Inter-rater reliability: two raters rate dataset A and two raters rate dataset B how to combine it into one number?

1 vote
1 answer
45 views

validate correctness of prediction

1 vote
0 answers
94 views

Correlation test over time

  • 11
1 vote
0 answers
903 views

How to validate a new measurement tool?

1 vote
2 answers
2k views

What is more important in a test set sensitivity/specificity or ROC AUC

0 votes
0 answers
137 views

Observer agreement in multilevel binary data

  • 245
0 votes
0 answers
1k views

Agreement statistics for confusion matrix with nominal categories

  • 7,863
0 votes
1 answer
637 views

Calculation for inter-rater reliability where raters don't overlap and different number per candidate?

  • 55
0 votes
0 answers
60 views

Is it legitimate to compare the accuracy between SVM models?