Quinlan, J. R., & Rivest, R. L. (1989). Inferring decision trees using the Minimum Description Length Principle. Information and Computation, 80, 227-248.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... Minimum Descrip tion Length (MDL) principle [20] The MDL principle essentially provides an information theoretic definition of the optimal theory model that can be inferred from a set of data examples; and it has been applied in a variety of problems (e.g. constructing decision trees [19], learning common patterns in a set of strings [18] inferring DTDs from a collection of XML data [13] An important observation made in [13] is that for two given REs R i and R j , if R i defines a larger language than R j (i.e. R i is less precise than R j with respect to the same collection ....

J.R. Quinlan and R.L. Rivest. Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, 80:227--248, 1989.

....(DAC) also known as recursive partitioning, is a technique that generates hierarchically organized rule sets (decision trees) In this work, DAC is combined with the information gain criterion [9] for selecting branching features. Furthermore, the minimum description length (MDL) criterion [16], modified to handle numerical attributes effectively, is used to avoid over fitting. This method, referred to as DAC MDL, is preferred instead of splitting the training data into one grow and one prune set. Splitting into grow and prune set for this data is more likely to result in highly ....

....set. This strategy is combined with incremental reduced error pruning [19] where each clause immediately after its generation is pruned back to the best ancestor. The criterion for choosing the best ancestor is to select the most compressive rule using an MDL coding scheme similar to the one in [16] but adapted to the single rule case. The method generates an unordered set of rules, in contrast to generating a decision list [10] This means that rules are generated independently for each class, and any conflicts due to overlapping rules are resolved during classification by using the nave ....

Quinlan and Rivest (1989) "Inferring Decision Trees Using the Minimum Description Length Principle", Information and Computation 80(3) (1989) 227-248

....On the other hand, the design of real learning systems seems to be dominated by ad hoc trial and error methods. Applications of these recent theoretical results to real world learning system design are scarce and far between. One exception is the elegant paper by Quinlan and Rivest, [14]. In a companion paper [26] we develop the theory and mathematical validation of the MDL principle [1,28,29] based on ultimate data compression up to the Kolmogorov complexity. Our purpose here is trying to bring theory and practice together by testing the theory on simple real applications. We ....

J. Quinlan, R. Rivest, Inferring decision trees using the minimum description length principle, Inform. Computation 80 (1989) 227--248.

....is partitioned 2. number of internal nodes, #I # #, 3. average depth of the tree, that is, average number of tests required before taking a decision, 4. topological relevance [8] More complex measures are obtained by encoding decision trees and the data not explained by the decision trees [9, 10]. Unfortunately, no universal code is known, and, as Wallace and Patrick pointed out, the adoption of a particular code for encoding the tree induced from the data is equivalent to accepting a certain prior probability distribution over the set of possible [trees] Obviously, an analogous ....

Quinlan JR, Rivest LR. Inferring decision trees using the minimum description trees using the minimum description length principle. Information and Computation 1989; 80:227}248.

.... languages, we present an alternative metric that attempts to address some of the shortcomings of the Max Count measure and is based on using Rissanen s Minimum Description Length (MDL) principle [21] The MDL principle has been applied in a variety of problems (e.g. constructing decision trees [20], learning common patterns in a set of strings [18] inferring DTDs from a collection of XML data [13] An important observation made in [13] is that for two given REs defines a larger language than (i.e. is less precise than with respect to the same collection of input data ....

J.R. Quinlan and R.L. Rivest. Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, 80:227--248, 1989.

....that almost all the records arriving at every node take a single categorical value (a numeric value close to the average) of the objective attribute with a high probability, and hence the single value (the average) could be a good predictor of the objective attribute. Making decision trees [9] [11] and regression trees [2] has been a traditional research topic in the field of machine learning and artificial intelligence. Recently the e#cient construction of decision trees and regression trees from large databases has been addressed and well studied among the database community and the KDD ....

J. R. Quinlan and R. L. Rivest. Inferring decision trees using minimum description length principle. Information and Computation, 80:227--248, 1989.

....c or speci c to general searches are used, our approach follows a short to long search. Hence, the splitting criterion is based on the Minimum Description Length (MDL) principle. The MDL principle has been previously used in the induction of decision trees in the post pruning phase [11, 21]. Also, the MDL principle has been used as a stopping criterion (pre pruning) 15, 17] as a measure for globally evaluating discretizations of continuous attributes [14] and for restructuring decision trees [13] In our approach, the MDL principle is used at the generation phase which is justi ....

J. R. Quinlan and R. L. Rivest. Inferring Decision Trees Using The Minimum Description Length Principle. Information and Computation, 80:227-248, 1989.

....of a theory may be so simple that it will probably model the data badly. Thus, when evaluating hypotheses it is often the case that greater accuracy is achieved in the description of new data with a small but imperfect theory, than 37 with a theory that perfectly describes the known data [Quinlan and Rivest 1989]. The MDL Principle can be incorporated into Bayesian Learning, which is employed as the basis of some language learning systems. Bayes theorem is defined as: P (H D) P (H)P (D H) P (D) where: the term P(H) is the prior probability, which represents the learner s belief in the ....

Quinlan, J.R. and Rivest, R.L. Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, v.80, n. 3, p. 227-248, 1989.

....running one of these batch algorithms to build a new network structure. Most of the Bayesian network learning algorithms learn conditional probabil ity tables rather than trees. The exception to this is [45] In that work, conditional probability trees are learned directly using techniques from [87] which, as the au thors show, makes learning more efficient because of the reduction in parameters that need observations. Unfortunately, there is one major problem with learning the structure in this way, and it is a problem we do not yet have a solution for. All the Bayesian network learning ....

J. R. Quinlan and R. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227 248, 1989.

....induction approaches. Our method uses a heuristic based on the Minimum Description Length (MDL) principle [21] Hence, the decision tree is built in a short to long way. The MDL principle has been previously used in the induction of decision trees but just within the post pruning phase [11, 20]. Also, the MDL principle has been used as a stopping criterion (pre pruning) 17] as a measure for globally evaluating discretisations of continuous attributes [15] and for restructuring decision trees [14] In our approach, the MDL principle is used at the generation phase which is justi ed ....

....that fall into that node (i.e. E ) When pruning is activated, the stopping criterion is given by the pruning criterion that we describe next. The MDL criterion has been used for pre pruning and post pruning decision trees. Usually, a predictive MDL criterion has been applied for this purpose [20]. Note that exceptions are much costlier in the case of predictive MDL criterion, because there is a great di erence from regular examples (no extra bits are needed) and exceptions (the arguments and class need be coded) This means that the predictive MDL criterion would prune too late in many ....

J. R. Quinlan and R. L. Rivest. Inferring Decision Trees Using The Minimum Description Length Principle. Information and Computation, 80:227-248, 1989.

....) DL # : 1 , Tm 2 , true#, PR # : 1 , Prm 2 , Pr(S # )# # # = #(ex 1 , DL # ) if (# # # ) then goto noprune DL : DL # , PR : PR # , m : m 1, # = # # noprune: output DL, PR Figure 5. 1: Our algorithm for searching decision lists from examples techniques [15, 21, 29]. This principle selects the best model from a given set of candidate stochastic models and is stated as select the model in the observed data that permits the shortest encoding both of the observations and the model. Grounded in this principle, we formalize a set of stochastic models ....

....the function f 2 (A(#) we employ the model that permits the shortest code length. We here present the coding scheme for the regression trees, that are used for representing the target function. The code length for a structure of the regression tree equals the total number of nodes. The article [21] presents a full explanation of the code length and of the coding scheme for the tree. For each non terminal node, a threshold and an index at the node must be encoded. The threshold is encoded in the same scheme as that used for the threshold of the decision lists in Appendix A, and the code ....

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227--248, 1989.

....of the interpretation. While Collins has devised several heuristic methods for estimating the probability model, further investigation into learning methods for this model still appears necessary. Magerman (1995) proposes a new parsing approach based on probabilistic decision tree models (Quinlan and Rivest, 1989; Yamanishi, 1992a) to replace conventional context free parsing. His method uses decision tree models to construct parse trees in a bottom up and left to right fashion. A decision might be made, for example, to create a new parse tree node, and conditions for making that decision might be, for ....

....respect to given data is that which requires the shortest code length in bits for encoding the model itself and the data observed through it. In this section, we will consider the basic concept of MDL and, in particular how to calculate description length. Interested readers are referred to (Quinlan and Rivest, 1989; Yamanishi, 1992a; Yamanishi, 1992b; Han and Kobayashi, 1994) for an introduction to MDL. 2.7.1 Basics of Information Theory IID process Suppose that a data sequence (or a sequence of symbols) is independently generated according to a discrete probability distribution P (X) 2.6) where ....

Quinlan, J. Ross and Ronald L. Rivest. 1989. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227--248.

.... Minimum Descrip tion Length (MDL) principle [20] The MDL principle essentially provides an information theoretic definition of the optimal theory model that can be inferred from a set of data examples; and it has been applied in a variety of problems (e.g. constructing decision trees [19], learning common patterns in a set of strings [18] inferring DTDs from a collection of XML data [13] An important observation made in [13] is that for two given REs # , if # defines a larger language than # (i.e. # is less precise # with respect to the same collection of input data ....

J.R. Quinlan and R.L. Rivest. Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, 80:227--248, 1989.

....of removal is repeatedly performed until the rule could not be generalized further. After all the initial rules are generalized, they are grouped into rule sets corresponding to the classes respectively. All rule sets are polished with the help of the Minimum Description Length (MDL) Principle [21] so that rules that do not contribute to the accuracy of a rule set are removed. Then, the rule sets are sorted according to the ascending order of their false positive error rates. Finally, a default rule is created for dealing with instances that are not covered by any of the generated rules. ....

J. R. Quinlan and R. L. Rivest, "Inferring decision trees using the minimum description length principle," Information and Computation, vol. 80, no.3, pp.227-248, 1989.

....is conserved by pursuing only the most highly rated partial solutions. In the implementation, ratings were calculated based upon a minimum description length (MDL) parameter which estimated the total length of a code required to represent the input signal using the model implicit in the hypothesis [18]. Formally equivalent to Bayesian analysis, MDL permits the integration of model complexity, model parameterization complexity, and goodness of fit into a single number and comprised a consistent theoretical basis for ratings assigned to otherwise disparate hypotheses. Hz Bad dog loo ....

J.R. Quinlan, R. L. Rivest. "Inferring decision trees using the Minimum Description Length principle," Information and Computation 80(30), 1989, 227-248.

....write on p. 115: Note that it was not even known whether nding the smallest decision tree is NP hard. It was only known that nding a decision tree with the minimum external path length is NP hard [9, 4] Quinlan Rivest look at a more general notion of decision trees than our notion in [20]. A set of objects each having a set of attributes (not necessarily two valued) is given. Each internal node tests an attribute and leaves are labelled negative or positive. Reading objects as elements in f0; 1g attributes as variables, we see that our notion of decision trees falls under ....

....(not necessarily two valued) is given. Each internal node tests an attribute and leaves are labelled negative or positive. Reading objects as elements in f0; 1g attributes as variables, we see that our notion of decision trees falls under Quinlan Rivest s more general notion. p. 242 in [20]: It is probably dicult to compute the best decision tree under our measure . Hya l and Rivest (1976) prove that constructing an optimal binary decision tree is NP complete when the cost of a tree is its external path length; it may be possible to modify this proof to handle the current ....

[Article contains additional citation context not shown here]

Quinlan, J.R., and Rivest, R.L., \Inferring Decision Trees Using the Minimum Description Length Principle", in: Information and Computation 80 (1989) 227-248.

....S#T# # # S#T### # DL ## #T # # : # T### # true#, PR ## #Pr # # : Pr### # Pr#S # #ex # # DL if ( # ) then goto noprune DL ## DL , PR## PR , m ## m# #, # noprune: output DL, PR Figure 5. 1: Our algorithm for searching decision lists from examples techniques [15, 21, 29]. This principle selects the best model from a given set of candidate stochastic models and is stated as select the model in the observed data that permits the shortest encoding both of the observations and the model. Grounded in this principle, we formalize a set of stochastic models ....

....the function f # #A# ##, we employ the model that permits the shortest code length. We here present the coding scheme for the regression trees, that are used for representing the target function. The code length for a structure of the regression tree equals the total number of nodes. The article [21] presents a full explanation of the code length and of the coding scheme for the tree. For each non terminal node, a threshold and an index at the node must be encoded. The threshold is encoded in the same scheme as that used for the threshold of the decision lists in Appendix A, and the code ....

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227--248, 1989.

....a large grammar augmented with a compositional semantics [Grover et al., 1993] Again, this should lead to better results. Prior weighting. As is well known, MDL based learners sometimes improve from weighting the prior with respect to the likelihood. Schemes, such as Quinlan and Rivest s [Quinlan and Rivest, 1989], fall outside of the coding framework and (e ectively) replicate the training set. We intend to pursue encoding based schemes that achieve the same purpose. 4 We have now built started experimenting with Random Field models [Osborne, 2000] Preliminary results show improvements over parse ....

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227-248, 1989.

....complexity to choose one of the h i output by G. In the literature can be found a number of methods for handling this trade off; these include cross validation [2, 11] using a separate hold out set on which to test the sequence of hypotheses produced [3] the minimum description length principle [7], Vapnik s structural risk minimization [9, 10] and heuristic estimates of the actual error of the best hypothesis found for a given size bound [6] Any of these methods can be used in combination with algorithm G. BBG uses the last approach mentioned. The various h i are compared according to a ....

Quinlan, J. R., & Rivest, R. L. (1989.) Inferring decision trees using the Minimum Description Length Principle. Information and Computation 80, 227--248.

....the methods and parameters that were used in the experiments. 3.2.1 Decision tree induction using MDL A method for decision tree induction was defined by choosing the divide and conquer strategy together with the information gain criterion. An MDL criterion (a version of the criterion in [25] that has been modified to handle numerical attributes effectively) was chosen as a means to avoid over fitting. This method was selected in favor of splitting the training data into a grow and a prune set, which due to the limited size of the data set would lead to highly variable decision ....

Quinlan J.R and Rivest R.L, "Inferring Decision Trees Using the Minimum Description Length Principle", Information and Computation 80(3) (1989) 227-248

....have been proposed for classification, such as decision trees, neural networks, Bayesian belief networks, fuzzy sets, and generic models. Among these models, decision trees are widely used for classification. We focus on decision tree induction in this paper. ID3 (and its variants such as C4.5) [1, 2] and CART [4] are among the best known classifiers that use decision trees. Other decision tree classifiers include Interval Classifier [3] and SPRINT [3, 5] which concentrate on making it possible to mine databases that do not fit in main memory by only requiring sequential scans of the data. ....

....attribute; the others are non class attributes. We store the decision path for each node. For example, in the decision tree below (Figure 2) the decision path for node N09 is Band2, value 0011, Band3, value 1000 . We use RC to denote the root count of a P tree, given node N s decision path B[1], V[1] B[2] V[2] B[t] V[t] let P tree P=P B[1] v[1] P B[2] v[2] P B[t] v[t] B2 0010 0011 0111 1010 1011 B1 B3 B1 B1 B1 0111 0100 1000 0011 1111 0010 B1 B1 0111 0011 We can calculate node N s information I(P) through # = n p p P I 1 2 log ) where p i = ....

[Article contains additional citation context not shown here]

J. R. Quinlan and R. L. Riverst, "Inferring decision trees using the minimum description length principle", Information and Computation, 80, 227-248, 1989.

....Kehagias thanks V.G. Kaburlasos for introducing him to lattice theory in general, and to lattice hyperbox operations in particular. He thanks Karen VanDyck, V.G. Kaburlasos and V. Petridis for many stimulating conversations. 1 by structures such as classification, regression and decision trees [3, 23] as well as in neural [4, 18] and fuzzy [12] extensions of such structures. The use of hyperboxes is implicit in the above cited examples; there are also cases [5, 6, 7, 8, 28, 29] where the hyperbox terminology is used explicitly. Similarly, hyperbox terminology appears in the theory of ....

J.R. Quinlan and R.L. Rivest, Inferring decision trees using the minimum description length principle, Inf. and Comp., vol.80, p.227-248.

.... guarantees to find the best (according to the misclassification cost) pruned decision tree T r , r # 1, 2, R , of the original tree T 0 of a specific size (as dictated by the complexity parameter) An alternative pruning method using Rissanen s minimum description length is described in (Quinlan and Rivest, 1989). 3.2. Pruning meta classifiers The post training pruning algorithm employs the minimal cost complexity method as a means to reduce the size (number of base classifiers) of the meta classifiers. 3 Estimated over a separate pruning subset of the training set or using cross validation methods. ....

Quinlan, R. and Rivest, R. (1989), `Inferring decision trees using the minimum description length princliple', Information and Computation 80, 227--248.

.... guarantees to find the best (according to the misclassification cost) pruned decision tree Tr , r # 1, 2, R , of the original tree T0 of a specific size (as dictated by the complexity parameter) An alternative pruning method using Rissanen s minimum description length is described in [35]. 3 Estimated over a separate pruning subset of the training set or using cross validation methods. Figure 1: The six steps of the Post Training Pruning Algorithm 3.2 Pruning meta classifiers The post training pruning algorithm employs the minimal cost complexity method as a means to reduce ....

R.J Quinlan and R.L. Rivest. Inferring decision trees using the minimum description length princliple. Information and Computation, 80:227--248, 1989.

....to predict the class label of unclassified data. Different models have been proposed for classification, such as decision tree induction, neural network, Bayesian, fuzzy set, nearest neighbor and so on. Among these models, decision tree induction is widely used for classification, such as ID3, C4.5[1,2], CART[4] Interval Classifier[3] SPRINT[3,5] and BOAT[6] Both association rule mining and classification have been applied in many fields. Remotely Sensed image data is one of the promising application areas since there are huge amount of image data. However, due to the large size of image ....

J.R.Quinlan,R.L.Riverst,"Inferringdecision trees using minimum description length principle", Information and Computation, 80, 227-248, 1989.

.... false A multiple decision tree for the only in useful bunch concept T Examples of the concept useful bunch : T1 (useful bunch) T2 (useless bunch) The MDL principle used to define an explicit measure of the complexity for encoding the training instances and the multiple decision tree (Quinlan and Rivest 1989) can be extended similarly. Based on the multiple entropy measure, ID3 M, C4.5 M have been built as multiple versions of the corresponding algorithms. Learning multi rules This section focuses on set of rules learners that are based on a coverage measurement. The growing procedure of the set of ....

Quinlan, J. R. and R. Rivest 1989. "Inferring Decision Trees Using the Minimum Description Length Principle." Information and Computation 80: 227-248.

....increasing amount and complexity of today s data, there is an urgent need to accelerate discovery of information in databases. In response to this need, numerous approaches have been developed for discovering concepts in databases using a linear, attribute value representation [1] 2] 3] 4] [5]. These approaches address issues of data relevance, missing data, noise, and utilization of domain knowledge. However, much of the data that is collected is structural in nature, or is composed of parts and relations between the parts. Hence, there exists a need for methods to analyze and ....

....DESCRIPTION LENGTH PRINCIPLE The minimum description length (MDL) principle introduced by Rissanen [17]states that the best theory to describe a set of data is a theory which minimizes the description length of the entire data set. The MDL principle has been used for decision tree induction [5], image processing [18] 19] 20] concept learning from relational data [21] and learning models of non homogeneous engineering domains [22] We demonstrate how the minimum description length principle can be used to discover substructures in complex data. In particular, a substructure is ....

[Article contains additional citation context not shown here]

J.R. Quinlan and R.L. Rivest, "Inferring Decision Trees Using the Minimum Description Length Principle," Information and Computation, vol. 80, pp. 227--248, 1989.

....decision trees carries through naturally to the scoring methods used to select the next rule to apply. Decision trees often select the split which most reduces either a diversity index or some measure based on the conditional entropy of the truth given the tree s predictions (Breiman et al. 1984; Quinlan and Rivest, 1989; Quinlan, 1993) Note that these metrics may select a split that does not change the score of the current 4 predictions against the truth, for instance by splitting a node in such a way that both children still have the same plurality class as the parent. Such a split may still make sense in ....

Quinlan, J. Ross and Ronald L. Rivest. 1989. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227-- 248.

....to noise and in such cases the proposed method is acceptable. Section 7.1 provides some examples. Now the next task is to construct a method for encoding the above tree structure. Various methods for encoding and optimizing classification trees are proposed in the literature; see, for example, Quinlan Rivest (1989), Wallace Patrick (1993) and Rissanen (1997) However, these methods are not suitable for our purposes, as they all assume that all internal nodes of a classification tree have two children. Note also that the tree structure implicitly defined by the wavelet recursive partitioning scheme of ....

Quinlan, J. R. & Rivest, R. L. (1989), `Inferring decision trees using the minimum description length principle', Information and Computation 80, 227--248.

....of each option. We are also interested in exploring di erent ways to avoid over tting through pruning and rules that decide when to stop splitting, such as the cost complexity pruning technique of Breiman [3] or the minimum description length approach suggested by Quinlan and Rivest [13]. However, since pruning takes place after the creation of the tree, and is not computationally intensive to bene t from parallel processing, we do not address the topic in this paper. 3.2 The Sapphire Decision Tree Design As explained in the previous section, we are interested in a decision ....

Quinlan, J. R., and Rivest, R. Inferring decision trees using the minimum description length principle. Information and Computation 80, 3 (1989), 227-248.

....data generally leads to decision trees with few covered cases per leaf, hence the problem of trading off bias and variance. This paper on decision tree pruning is manifestly incomplete: Space constraints obliged us to neglect several other pruning methods presented in the literature. One of them [27] is part of a decision tree induction process based on the minimum description length (MDL) principle [28] Given an efficient technique for encoding decision trees and exceptions, which are examples misclassified by the decision tree, the MDL principle states that the best decision tree is the ....

J.R. Quinlan and L.R. Rivest, "Inferring Decision Trees Using the Minimum Description Length Principle," Information and Computation, vol. 80, pp. 227-248, 1989.

No context found.

Quinlan, J. R., & Rivest, R. L. (1989). Inferring decision trees using the Minimum Description Length Principle. Information and Computation, 80, 227-248.

No context found.

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, pages 227--248, 1989.

No context found.

Quinlan, J. R. and Rivest, R. L. (1989). Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, 80:227--248.

No context found.

R. J. Quinlan and R. L. Rivest, "Inferring decision trees using the minimum description length principle," Information and Computation, vol. 80, pp. 227--248, 1989.

No context found.

J.R. Quinlan and R.L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227--248, 1989.

No context found.

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80(3):227--248, 1989.

No context found.

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the Minimum Description Length Principle. Inform. Comput., 80(3):227--248, March 1989. (An early version appeared as MIT LCS Technical report MIT/LCS/TM-339 (September 1987).).

No context found.

J.R. Quinlan, R.L. Rivest. Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation 80(3): 227-248, 1989.

No context found.

J. R. Quinlan and R. Rivest. Inferring Decision Trees using the Minimum Description Length Principle. Information and Computation, 80:227--248, 1989.

No context found.

J. Quinlan and R. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80:227-248, 1989.

No context found.

J. R. Quinlan and R. L. Rivest. Inferring decision trees using the minimum description length principle. Information and Computation, 80(3):227--248, 1989.

No context found.

J.R. Quinlan and R. Rivest, "Inferring Decision Trees Using the Minimum Description Length Principle", Information and Computation, Vol 80, pp.227-248, March, 1989

No context found.

J. R. Quinlan and R. L. Rivest. Inferring Decision Trees Using The Minimum Description Length Principle. Information and Computation, 80:227={248, 1989.

No context found.

Quinlan, J. & Rivest, R. (1989) Inferring decision trees using the minimum description length principle. Information and Computation 80, 227-248

No context found.

J.R. Quinlan and R.L. Rivest. "Inferring Decision Trees Using the Minimum Description Length Principle". Information and Computation, 80:227--248, 1989.

No context found.

Quinlan, J. R. and Rivest, R. L. (1989). Inferring Decision Trees Using the Minimum Description Length Principle. Information and Computation, 80:227--248.

No context found.

J.R. Quinlan, R.L. Rivest, Inferring decision trees using the Minimum Description Length principle, Information and Computation 80, 227--248 (1989).

No context found.

J. R. Quinlan & R. L. Rivest (1989.) Inferring decision trees using the Minimum Description Length Principle. Information and Computation 80, 227--248.

No context found.

BIBLIOGRAPHY 190 QUINLAN,J.ROSS,&RONALD L. RIVEST. 1989. Inferring decision trees using the minimum description length principle. Information and Computation 80.227--248.

First 50 documents  Next 50

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