C50 and machine learning

An existing data set, containing experts on whether contact lenses can be used for diagnostic records (from "Data Mining"), try to use r language to implement rules extraction.

Structure

> spectacle = Factor (Rep (Rep ("Myope", 4), Rep ("Hypermetrop", 3)), 3) > Age = Factor (C (Rep ("Young", 8), Rep (" Pre-presbyopic ", 8), Rep (" Presbyopic ", 8))) > spectacle = Factor (Rep (Rep (" Myope ", 4), Rep (" Hypermetrop ", 4)), 3)) > Astimatism = Factor (Rep (C ("No", "no", "yes", "yes"), 6)) > tear = Factor (Rep (c ("reduced", "normal")) > recommended = Factor (C ("None", "soft", "none", "hard", "none", "soft", "none", "hard", "none",                       "soft", "none", "hard", "None", "soft", "none", "none", "none", "none",                        "None", "hard", "none", "soft", "none", "none")) > DF <- Data.frame (age,spectacle,astimatism,tear,recommended)

Rule generation

> Model <-rpart (Formula = recommended ~.,data = DF2) > Summary (model) Call:rpart (formula = Recommended ~., data =      DF2) n= CP nsplit rel error xerror xstd1 0.2222222 0 1.0000000 1.000000 0.26352312 0.0100000   1 0.7777778 1.333333 0.2721655Variable importancetear Node number 1:24 observations, complexity param=0.2222222 Predicted Class=none expected loss=0.375 P (node) =1 class Counts:4 5 probabilities:0.167 0.625 0. 208 left son=2 (OBS) Right son=3 (OBS) Primary splits:tear splits as RL, improve=5.0833330, (0 miss ing) astimatism splits as RL, improve=1.7500000, (0 missing) age splits as RRL, improve=0.2916667, (0 Missing) spectacle splits as RL, improve=0.2500000, (0 missing) Node number 2:12 observations predicted Class=non E expected loss=0 P (node) =0.5 class counts:0 0 probabilities:0.000 1.000 0.000 node number 3:12 o Bservations predicted CLass=soft expected loss=0.5833333 P (node) =0.5 class Counts:4 3 5 probabilities:0.333 0.250 0.417  

Visualization of

> par (XPD = TRUE) > plot (model) > Text (model)

Statistical summarization of algorithm C5.0

Call:c5.0.formula (formula = Recommended ~., data = DF2) C5.0 [Release 2.07 GPL Edition]  Mon Mar 14:47:09------ -------------------------Class specified by attribute ' outcome ' Read cases (5 attributes) from Undefined.datadecision t Ree:tear = Reduced:none (tear) = Normal::...astimatism = No:soft (6/1)    astimatism = Yes:hard (6/2) Evaluation on t Raining data (cases):    decision Tree     ----------------    Size      Errors       3    3 (12.5%)   < <   (a)   (b)   (c)    <-classified as  ----  ----  ----     4                (a): Class hard     2     1    (b): Class none                 5    (c): Class Softattribute Usage:100.00%tear 50% astimatismtime:0.0 secs

It was found that the secretion of lacrimal glands was increased after the doctor's decision was worn.

C50 and machine learning