Machine learning app for pigmented skin lesion diagnosis prediction with HAM10000 dataset (https://www.nature.com/articles/sdata2018161).

Testing with unseen data (1992 cases), the statistical result is as follow:

Overall accuracy: 76.66% (1527/1992)

akiec sensitivity: 45.21%(33/(33+40))
akiec precision: 58.93% (33/(33+23))

bcc sensitivity: 55.86%(62/(62+49))
bcc precision: 63.27% (62/(62+36))

bkl sensitivity: 49.33%(110/(110+113))
bkl precision: 58.20% (110/(110+79))

df sensitivity: 31.82%(7/(7+15))
df precision: 58.33% (7/(7+5))

mel sensitivity: 32.75%(75/(75+154))
mel precision: 50.34% (75/(75+74))

nv sensitivity: 94.27%(1234/(1234+75))
nv precision: 83.49% (1234/(1234+244))

vasc sensitivity: 24.00%(6/(6+19))
vasc precision: 60.00% (6/(6+4))

*Abbreviations
akiec: actinic keratoses and intraepithelial carcinoma/bowen's disease
bcc: basal cell carcinoma
bkl: benign keratosis-like lesions (solar lentigines/seborrheic keratoses and lichen-planus like keratoses)
df: dermatofibroma
mel: melanoma
nv: melanocytic nevi
vasc: vascular lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhage)
Read more
Collapse
Loading…

What's New

Add calculation for top 3 accuracy.
Read more
Collapse

Additional Information

Updated
November 20, 2019
Size
5.5M
Installs
10+
Current Version
2.0.2
Requires Android
6.0 and up
Content Rating
Everyone
Permissions
Offered By
AlmanacSoft
Developer
9 Samakhee 60/4 Thasai, Muang, Nonthaburi, THAILAND 11000
©2020 GoogleSite Terms of ServicePrivacyDevelopersArtistsAbout Google|Location: United StatesLanguage: English (United States)
By purchasing this item, you are transacting with Google Payments and agreeing to the Google Payments Terms of Service and Privacy Notice.