J.R. Quinlan. Induction of Decision Trees. Machine Learning, Vol. 1, pp. 81--106, Kluwer Academic Publishers, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....frequency and information measures can be obtained from their discrete domains in a straightforward manner without prior partitioning. Usually, for symbolic data an explicit ordering scheme is given, like rule inference, which can be represented as data driven decision tree with labeled leaves [1, 2]. Classic programs like FOIL [3] and GOLEM [4] both provide an induction of Horn clauses from data, but the domain of ILP has been extended to dynamic hypothesis generation [5] learning recursive logic [6] and program synthesis [7, 8] Soft representations are especially suitable for real value ....

....c.hypo p.hypo normal c.hypo p.hypo Fig. 1. Data preparation. Left: scatter plot of original data, hypothyroid features TSH versus T3. Right: same data after logarithmic and z transform. Normalization is suggested in the final preparation step. Rescaling the data components to the interval [0, 1] is not tolerant to outliers, therefore, mean subtraction followed by rescaling the standard deviation to 1 is preferred, which is in statistics is known as z transform. The mean and standard deviation for each dimension are saved for being applied to the test sets and to new data, and for ....

J.R. Quinlan. Induction of Decision Trees. Machine Learning, Vol. 1, pp. 81--106, Kluwer Academic Publishers, 1986.

....meaning that most users will only be able to capture top level splits. As such, although no extensive experimental comparisons have been ca, tied out till now 5, it has been taken for grated that end cut splits are undesirable, ad most existing tree based systems (e.g. CART [2] THAID [ or C4. [7]) have some mechanism for avoiding them. However, if the drawbacks in terms of user expectations re irrefutable, as we will see in Section 3 the drawbacks of end cut splits in terms of predictive accuracy ae not so clea at all in the case of least squares regression trees. 2 In [8] a similar ....

J. Quinlan. Cd.5: programs for machine learning. Kluwer Academic Publishers, 1993.

....the average predictive accuracy for this episode is greater or equal than a predefined threshold t . The threshold for the episode selecting criterion in step 2) is setup to 6 . 0 = t in all experiments. The learning algorithm used in steps 1) and 3) is Induction on Decision Tree (aka ID3) [14]. The reason for selecting this algorithm was that this algorithm generates explicit user profiles, which is an important aspect for user modeling. This was the reason for this algorithm to be used in CAP as well. Prediction task CAP ID3 FM Context Learning Location 64 58 67 Duration 63 71 79 ....

Quinlan, R.: Induction of Decision Trees. Machine Learning 1, Kluwer Academic Publishers (1986), 81-106.

....classifying examples. In this task, the system is given a list of examples in the form of a set of predicates followed by a classification symbol. The goal is to predict what symbol is correct for a given set of predicates. Many approaches have been taken to solve this problem. The algorithm ID3 [36] builds a decision tree using the predicates as choice points. Another solution to the problem was Mitchell s concept space algorithm [25] A concept space represents the set of all solutions as a graph and each positive and negative example cuts off portions of the graph until one is left with ....

J.R. Quinlan. Induction of Decision Trees. Machine Learning, Kluwer Academic Publishers, Boston, Vol. 1, 1986, pp 81-106.

....it that explains why each event occurred. Gamut uses this to infer where the software author means to create new actions, conditions, and loops. The second form of induction is decision trees, which Gamut uses to learn the complex relationships between modes. A decision tree algorithm like ID3 [16] uses statistical measurements to map the values of a set of attributes into a domain of concept names. In Gamut, the attributes are predicates on the state of the objects in the game, and the concepts are branches in a conditional statement. 4. Interaction Techniques Gamut s interaction ....

J. R. Quinlan. Induction of Decision Trees. Machine Learning, Kluwer Academic Publishers, Boston, Vol. 1, 1986, pp 81-106.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC