Reading decision trees - artificial intelligence, Computer Engineering

Reading Decision Trees

There is a link between decision tree representation and logical representations, which may be exploited to form it more easy  to understand and learned decision trees. If we assume about it, each decision tree is actually a disjunction of implications (if ... then statements), and the implications are Horn clauses: a conjunction of literals implying a single literal. In the above tree, we may notice this by reading from the root node to each leaf node:

If the parents are tripping, then go to the cinema

or

If the parents are not tripping and it is sunny, then go fortennis

or

If the parents are not tripping and it is windy and you're wealthy, then go for shopping

or

If the parents are not tripping and it is windy and you're poor, then go to cinema

or

If the parents are not tripping and it is rainy, then stay inside.

Of course, this is only a re-statement of the actual mental decision making method we defined. Remember, however, that we will be programming an agent to learn decision trees from example, so this sort of conditions will not happen as we will start with just example conditions. It will therefore be important for us to be able to read the decision tree the agent advises.

Decision trees don't have to be representations of decision forming methods, and they can equally use to categorization problems. If we phrase the above question slightly in different way, we may see this: instead of saying that we need to represent a decision method for what to do on a weekend, we might ask what sort of weekend this is: is it a weekend where we play tennis, or one where we go for shopping, or one where we watch a film, or one where we stay inside? For other example, we can refer back to the animals example from the previous lecture:  in that  case, we wished to categories what group an animal was (mammal, fish, reptile, bird) using physical features (whether it lays eggs, number of legs, etc.). This could simply be phrased as a question of learning a decision tree to choose which category a given animal is in, e.g., if it lays eggs and is homoeothermic, then it's a bird, and so on...

Posted Date: 10/3/2012 1:07:40 AM | Location : United States







Related Discussions:- Reading decision trees - artificial intelligence, Assignment Help, Ask Question on Reading decision trees - artificial intelligence, Get Answer, Expert's Help, Reading decision trees - artificial intelligence Discussions

Write discussion on Reading decision trees - artificial intelligence
Your posts are moderated
Related Questions
How aliases are used in DNS? Explain. CNAME entries are analogous to a symbolic link in a file system- the entry gives an alias for other DNS entry. Corporation of expertsmind

Consider one versus the rest voting used for classifier with three classes {a, b, c}. Given a row of data denoted as x0 suppose that the classifier for a versus the rest predicts t

What is a binary tree? A binary tree 'T' is described as A finite set of elements, known as nodes, such that: either (i) or (ii) happens: (i) T is empty (known as the 'nu

Eight-stage process - Conjunctive normal forms: Hence we notice the following eight-stage process converts any sentence with CNF as:  1.  Eliminate all arrow connectives by

Massively Parallel System Refers to a parallel computer system having a great number of processors. The number in a great number of keeps increasing and currently it means more

Write  a  verilog  code  to  swap  contents  of  two registers  with  and without  a  temporary register? With temp reg : always @ (posedge clock) begin temp=b; b

Program Level This is usually the dependability of the operating system, which runs processes simultaneously. Different programs are evidently independent of each other. So par

Q. What is Hard Drive Interface? Secondary storage devices need a controller to proceed as an intermediary between device and rest of the computer system. On some computers the

A vertical flashing line that permits the user where text will be inserted.

Attributes are declarative tags in code that put in additional metadata into an assembly.