Torgo, L. and Gama, I. Regression Using Classification Algorithms. Intelligent Data Analysis, 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to each URL is done according to a sum of the discounted number of rewards that are obtainable from a URL in the future. A more immediate strategy does not consider the future rewards in Q value computation. Mapping from text to Q value is done by casting this regression problem as classification [62], discretizing hyperlink Q values and training a naive Bayes classifier on the corresponding neighborhood text. Figure 12 shows the pseudo code of mapping text to Q value in the training phase. Figure 13 shows the pseudo code of mapping from neighborhood text of URLs to Q Value in training data. ....

L.Torgo, J. Gama, Regression using classification algorithms, Intelligent Data Analysis 1(4), 1997.

....learning spider, we learn a mapping from the text in the neighborhood of a hyperlink to the expected (discounted) number of relevant pages that can be found as a result of following that hyperlink. The mapping from the text to a scalar is performed by casting regression as classification [ Torgo and Gama, 1997 ] Specifically, we discretize the scalar values into a finite number of bins, and use naive Bayes to classify the text into a corresponding finite number of classes; the value assigned to a particular hyperlink is a weighted average of the values of the top ranked bins. Our research in e#cient ....

....function that maps hyperlinks to scalar values. The mapping should be e#cient, and should generalize to future, unseen hyperlinks. We represent the value function using a collection of naive Bayes text classifiers, performing the mapping by casting this regression problem as classification [ Torgo and Gama, 1997 ] We discretize the discounted sum of future reward values of our training data into bins, place the text in the neighborhood of the hyperlinks into the bin corresponding to their Q values, and use the hyperlinks neighborhood text as training documents for a naive Bayes text classifier. For ....

Luis Torgo and Joao Gama. Regression using classification algorithms. Intelligent Data Analysis, 1(4), 1997.

....in this paper show that using a more restricted set of neighborhood text does not enhance spidering performance. Now using these features we must define a regression function, F , that maps text to real valued Q values. We perform the mapping by casting this regression problem as classification (Torgo Gama, 1997). We discretize the discounted sum of future reward values of our training data into bins and treat each bin as a class. For each hyperlink, we calculate the probabilistic class membership of each bin using naive Bayes (which is described in section 5.2.1) Then the Q value of a new, unseen ....

Torgo, L., & Gama, J. (1997). Regression using classification algorithms.

....the words in the neighborhood of the hyperlink corresponding to each action. Thus our Q function becomes a mapping from a bag of words to a scalar. We represent the mapping using a collection of naive Bayes text classifiers (see Section 3. 1) and cast this regression problem as classification [ Torgo and Gama, 1997 ] We discretize the discounted sum of future reward values of our training data into bins, place each hyperlink into the bin corresponding to its Q value (see below) and use the text in the hyperlink s anchor and surrounding page as training data for the classifier. At test time, the reward ....

L. Torgo and J. Gama. Regression using classification algorithms. Intelligent Data Analysis, 1(4), 1997.

....the reward function, R, are known, and we learn the Q function by dynamic programming in the original, uncollapsed state space. We represent the mapping using a collection of naive Bayes text classifiers (see Section 4. 2) We perform the mapping by casting this regression problem as classification (Torgo Gama 1997). We discretize the discounted sum of future reward values of our training data into ten bins, place the hyperlinks into the bin corresponding to their Q values as calculated above, and use the hyperlinks text as training data for a naive Bayes text classifier. For the anchor text of each ....

Torgo, L., and Gama, J. 1997. Regression using classification algorithms. Intelligent Data Analysis 1(4).

....learning spider, we learn a mapping from the text in the neighborhood of a hyperlink to the expected (discounted) number of relevant pages that can be found as a result of following that hyperlink. The mapping from the text to a scalar is performed by casting regression as classification [ Torgo and Gama, 1997 ] Specifically, we discretize the scalar values into a finite number of bins, and use naive Bayes to classify the text into a corresponding finite number of classes; the value assigned to a particular hyperlink is a weighted average of the values of the top ranked bins. Our research in efficient ....

....function that maps hyperlinks to scalar values. The mapping should be efficient, and should generalize to future, unseen hyperlinks. We represent the value function using a collection of naive Bayes text classifiers, performing the mapping by casting this regression problem as classification [ Torgo and Gama, 1997 ] We discretize the discounted sum of future reward values of our training data into bins, place the text in the neighborhood of the hyperlinks into the bin corresponding to their Q values, and use the hyperlinks neighborhood text as training documents for a naive Bayes text classifier. For ....

Luis Torgo and Joao Gama. Regression using classification algorithms. Intelligent Data Analysis, 1(4), 1997.

....the reward function, R, are known, and we learn the Q function by dynamic programming in the original, uncollapsed state space. We represent the mapping using a collection of naive Bayes text classifiers (see Section 5. 2) We perform the mapping by casting this regression problem as classification (Torgo Gama 1997). We discretize the discounted sum of future reward values of our training data into bins, place the hyperlinks into the bin corresponding to their Q values by dynamic programming, and use the hyperlinks neighborhood text as training data for a naive Bayes text classifier. We define a hyperlink s ....

Torgo, L., and Gama, J. 1997. Regression using classification algorithms. Intelligent Data Analysis 1(4).

No context found.

Torgo, L. and Gama, I. Regression Using Classification Algorithms. Intelligent Data Analysis, 1997.

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