Learning Profile
The learning profile is a deeper view of the patterns the system has extracted from your feedback history. It shows specific item combinations, occasion-level patterns, and weather preferences.
Best Combinations
A list of item pairs with high success rates in your outfit history. These are pairs where both items appeared together in accepted, positively-rated outfits more often than in rejected ones.
Each combination card shows:
- Thumbnails of the two items.
- Success rate: percentage of outfits containing this pair that were accepted.
- Times paired: how often these items have appeared together.
This is different from Smart Pairings, which shows general item combinations. The learning profile focuses on pairs that have personal significance to you based on your feedback history.
Occasion Patterns
A breakdown of your suggestion performance per occasion:
| Column | Description |
|---|---|
| Occasion | Casual, Office, Formal, Date, Sporty, Outdoor |
| Acceptance rate | How often you accepted suggestions for this occasion |
| Preferred colors | Colors that appeared most in accepted outfits for this occasion |
| Top items | Items that appeared most frequently in accepted outfits for this occasion |
Low acceptance rates for specific occasions often indicate the AI’s understanding of that occasion for your wardrobe needs improvement. Adding more feedback for that occasion (by generating and rating suggestions) fixes it over time.
Weather Preferences
Your preference patterns broken down by temperature band:
| Band | Description |
|---|---|
| Cold (below 10°C) | Average layer count in accepted outfits, preferred outerwear types |
| Cool (10-16°C) | Layering behavior, jacket vs. no jacket preference |
| Mild (16-22°C) | Light layer preferences |
| Hot (above 22°C) | Preference for minimal layers, fabric weight patterns |
This section helps you understand your personal temperature thresholds and whether they’ve been captured correctly by the system.
Preference Suggestions
Based on your feedback patterns, the system may suggest adjustments to your manually configured preferences:
- Items to add to favorites (based on wear frequency and acceptance patterns).
- Colors to add to your “avoid” list (based on consistent rejection of items in those colors).
- Style slider adjustments (if your observed behavior diverges from your configured sliders).
These are suggestions, not automatic changes. Click “Apply” on any suggestion to update your preferences, or dismiss it.
Recomputing the Profile
Click “Recompute Profile” to trigger a full refresh of all learning profile data. This re-runs all analysis from your feedback history. Use it after a period of high feedback activity, after significant wardrobe changes, or if the profile data feels stale.
Mobile: The learning profile is available in Settings > Learning Profile on mobile. Data is identical to the web view.