ID3 Algorithm Amrit Gurung
Classification Library System Organise according to special characteristics Faster retrieval New items sorted easily Related items clustered together
ID3 Algorithm Ross Quinlan Mathematical algorithm Decision tree Fixed examples Entropy Gain
ID3 Algorithm Entropy (randomness) Entropy(S) = S -p(I) log2 p(I) Range 0-1 Gain: Attribute property Gain(S, A) = Entropy(S) - S ((|S v | / |S|) * Entropy(S v ))
ID3 Algorithm Example NameHairHeightWeightLotionResult Sarahblondeaveragelightnosunburned Danablondetallaverageyesnone Alexbrownshortaverageyesnone Annieblondeshortaveragenosunburned Emilyredaverageheavynosunburned Petebrowntallheavynonone Johnbrownaverageheavynonone Katieblondeshortlightyesnone
ID3 Algorithm Example AttributeEntropy Hair Color0.50 Height0.69 Weight0.94 Lotion0.61
ID3 Algorithm Example AttributeEntropy Height0.50 Weight1.00 Lotion0.00
Limitations of ID3 Algorithm Considers only one attribute to create nodes Numerous trees needed for continuous data Over classification for small data
ID3 Algorithm Advantages Predict new data Training set is used to create rules for predicting