3 4 b, b, © Module-2 What is a decision tree & discuss the use of decision tree for classification purpose with an example (08 Marks) Write and explain decision tre for the following transactions: (08 Marks) “Tid Refund | Mactal status { Taxable income | Cheat 1] Yes_|_ Single 15K | No 2 [No | Married 100K [No 30[ No | Single 70K No’ [Wes | Married 120K | No 3_[-No | Divorced 95K Yes | No | Married ok No. 7_|Yes_[ Divorced 220K [No [No | Single SK Yes ‘9~| No | — Married 75K No. T0[ No | — Single 20K Yes OR For the transactions shown in the table compute the following (Entropy of the collection of transaction records of the table with respect to classification, Gi). What are the information gain of a and ay relative to the transactions of the table? (08 Marks) Tnsiance [7 a |T a |T Target cass | + Discuss the decision learning algorithm. (04 Marks) List the issues of decision tree learning. (04 Marks) lal a)=|>) + lalS| TR TIS a)=]o Module-2 Discuss the two approaches to prevent over fitting the data, (08 Marks) Consider the following set of training examples: Instance | Classification | ay 1 1 r 0 1 oO -|-lele[=|-le 1 T 0 0 0 on ) (i) What is the entropy of this collection of taining examples with respect to the target function classification? (Gi) What is the information gain of a: relative to these training examples? (8 Marks) or Define decision tree. Construct the decision tree to represent the following Boolean functions: AaB i) AV [BAC] iil) AXORB (06 Marks) Write the 1D3 algorithm {06 Marks) What do you mean by gain and entropy? How itis used to build the decision tree. (04 Marks) Modute-2 Consrst decision ree forthe following dating 1D3 algorh [Day AT AZ | AS] Clasifeaion 1 Tue Hoc} HighT No c Hot | High No.(3. false Wor [High 4) False” Cool | Normal 5__False Cool | Normal s 6) True Cool] High ‘No 7 Tre Hot | High No B_ True Hot | Normal | Yes ‘9 False Cool | Normal | Yes TO False Cool | High No (16 Marks) oR 4 a. Explain the concept of decision tree learning. Discuss the necessary measure required to select the atributes for building a decision tree using ID3 algorithm, (08 tanks) bb. Discuss the issues of avoiding over fitting the data, handling continuous data and missing values in decision tree. (08 Marks)