Learn Before
Impact of Multi-Category Labeling on Column Summaries
Question: If a single misclassified example (such as Image #3) has both 'Great Cat' and 'Blurry' checked in an error-analysis spreadsheet, how does this multi-category association affect the sum of the category percentages at the bottom of the spreadsheet?
Sample answer: Because a single example is counted in multiple columns (like 'Great Cat' and 'Blurry'), the categories are not mutually exclusive. This double-counting across different categories causes the sum of the column percentages at the bottom of the spreadsheet to fail to add up to 100% (and typically exceed 100%).
Key points:
- The example is counted in multiple categories/columns simultaneously.
- The category percentages are not mutually exclusive.
- The sum of the percentages at the bottom of the spreadsheet will not add up to 100%.
Rubric: The student must state that the example is counted in multiple columns, making the categories non-mutually exclusive, which results in the column percentages summing to something other than (or greater than) 100%.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Why may the category percentages in an error analysis spreadsheet not add up to 100%?
In an error analysis spreadsheet, each misclassified example must belong to exactly one error category.
Because one misclassified example can be associated with _____ categories, the column percentages in an error analysis spreadsheet may not add up to 100%.
Match each error analysis spreadsheet concept to its correct description.
Order the steps for conducting error analysis on misclassified dev-set examples using a category spreadsheet.
In Machine Learning Yearning, Image #3 has both the 'Great Cat' and 'Blurry' columns checked. What concept does this directly illustrate?
If column percentages in an error analysis spreadsheet sum to more than 100%, it necessarily indicates a data entry mistake was made.
In Machine Learning Yearning's error analysis illustration, Image #3 has both the Great Cat and the _____ columns checked.
Match each observation about an error analysis spreadsheet to the implication it directly supports.
Order the reasoning steps to correctly interpret column percentages that sum to more than 100% in an error analysis spreadsheet.
Analyzing the Overlap of Error Categories in Spreadsheet Summaries
Evaluating Non-Exclusive Column Sums in a Cat Detector Spreadsheet
Impact of Multi-Category Labeling on Column Summaries