Yusuke Suzuki

Discovery of Frequent Tag Tree Patterns in Semistructured Web Documents

Many Web documents such as HTML files and XML files have no rigid structure and are called semistructured data. In general, such semistructuredWeb documents are represented by rooted trees with ordered children. We propose a new method for discovering frequent tree structured patterns in semistructured Web documents by using a tag tree pattern as a hypothesis. A tag tree pattern is an edge labeled tree with ordered children which has structured variables. An edge label is a tag or a keyword in such Web documents, and a variable can be substituted by an arbitrary tree. So a tag tree pattern is suited for representing tree structured patterns in such Web documents. First we show that it is hard to compute the optimum frequent tag tree pattern. So we present an algorithm for generating all maximally frequent tag tree patterns and give the correctness of it. Finally, we report some experimental results on our algorithm. Although this algorithm is not efficient, experiments show that we can extract characteristic tree structured patterns in those data.

Ordered Term Tree Languages which Are Polynomial Time Inductively Inferable from Positive Data

In the fields of data mining and knowledge discovery, many semistructured data such as HTML/XML files are represented by rooted trees t such that all children of each internal vertex of t are ordered and t has edge labels. In order to represent structural features common to such semistructured data, we propose a linear ordered term tree, which is a rooted tree pattern consisting of ordered tree structures and internal structured variables with distinct variable labels. For a set of edge labels Λ, let OTTΛ be the set of all linear ordered term trees. For a linear ordered term tree t in OTTΛ, the term tree language of t, denoted by LΛ (t), is the set of all ordered trees obtained from t by substituting arbitrary ordered trees for all variables in t. Given a set of ordered trees S, the minimal language problem for OTTLΛ = {LΛ (t) | t ∈ OTTΛ} is to find a linear ordered term tree t in OTTΛ such that LΛ (t) is minimal among all term tree languages which contain all ordered trees in S. We show that the class OTTLΛ is polynomial time inductively inferable from positive data, by giving a polynomial time algorithm for solving the minimal language problem for OTTLΛ.

Polynomial Time Inductive Inference of Ordered Tree Patterns with Internal Structured Variables from Positive Data

Tree structured data such as HTML/XML files are represented by rooted trees with ordered children and edge labels. As a representation of a tree structured pattern in such tree structured data, we propose an ordered tree pattern, called a term tree, which is a rooted tree pattern consisting of ordered children and internal structured variables. A term tree is a generalization of standard tree patterns representing first order terms in formal logic. For a set of edge labels ? and a term tree t, the term tree language of t, denoted by L?(t), is the set of all labeled trees which are obtained from a term tree t by substituting arbitrary labeled trees for all variables in t. In this paper, we propose polynomial time algorithms for the following two problems for two fundamental classes of term trees. The membership problem is, given a term tree t and a tree T, to decide whether or not L?(t) includes T. The minimal language problem is, given a set of labeled trees S, to find a term tree t such that L?(t) is minimal among all term tree languages which contain all trees in S. Then, by using these two algorithms, we show that the two classes of term trees are polynomial time inductively inferable from positive data.

Discovery of Maximally Frequent Tag Tree Patterns with Contractible Variables from Semistructured Documents

In order to extract meaningful and hidden knowledge from semistructured documents such as HTML or XML files, methods for discovering frequent patterns or common characteristics in semistructured documents have been more and more important. We propose new methods for discovering maximally frequent tree structured patterns in semistructured Web documents by using tag tree patterns as hypotheses. A tag tree pattern is an edge labeled tree which has ordered or unordered children and structured variables. An edge label is a tag or a keyword in such Web documents, and a variable can match an arbitrary subtree, which represents a field of a semistructured document. As a special case, a contractible variable can match an empty subtree, which represents a missing field in a semistructured document. Since semistructured documents have irregularities such as missing fields, a tag tree pattern with contractible variables is suited for representing tree structured patterns in such semistructured documents. First, we present an algorithm for generating all maximally frequent ordered tag tree patterns with contractible variables. Second, we give an algorithm for generating all maximally frequent unordered tag tree patterns with contractible variables.

A polynomial time matching algorithm of structured ordered tree patterns for data mining from semistructured data

Tree structured data such as HTML/XML files are represented by rooted trees with ordered children and edge labels. Knowledge representations for tree structured data are quite important to discover interesting features which such tree structured data have. In this paper, as a representation of structural features we propose a structured ordered tree pattern, called a term tree, which is a rooted tree pattern consisting of ordered children and structured variables. A variable in a term tree can be substituted by an arbitrary tree. Deciding whether or not each given tree structured data has structural features is a core problem for data mining of large tree structured data. We consider a problem of deciding whether or not a term tree t matches a tree T, that is, T is obtained from t by substituting some trees for variables in t. Such a problem is called a membership problem for t and T. Given a term tree t and a tree T, we present an O(nN) time algorithm of solving the membership problem for t and T, where n and N are the numbers of vertices in t and T, respectively. We also report some experiments on applying our matching algorithm to a collection of real Web documents.

Experimental Study of the Active Compensation to a Full-Size Separate-Shell Magnetic Shield

Shielding efficiency of a full-size separate-shell magnetic shield equipped with active compensation coils is presented. Its size is 135 cm in height, 100 cm in width, and 240 cm in depth. With respect to active compensation two control schemes, i.e., feedback and feedforward, are compared. The shielding efficiency is estimated from the measurements of the attenuation of environmental magnetic fields in two axial directions (horizontal and vertical). In the case of feedback, a reference sensor that measures the incoming magnetic fields is positioned inside the shield, whereas in the case of feedforward the reference sensor is positioned outside the shell. In both axial directions, the environmental magnetic noise can be reduced to less than 500 pT: the noise floor measured by a fluxgate magnetometer is approximately 10 pT/¿Hz in the frequency range from 10 to 100 Hz.

Polynomial time inductive inference of TTSP graph languages from positive data

Two-Terminal Series Parallel (TTSP, for short) graphs are used as data models in applications for electric networks and scheduling problems. We propose a TTSP term graph which is a TTSP graph having structured variables, that is, a graph pattern over a TTSP graph. Let

Acquisition of Characteristic Tree Patterns with VLDC's by Genetic Programming and Edit Distance

\mathcal{TG_{TTSP}}

Efficient Learning of Unlabeled Term Trees with Contractible Variables from Positive Data

be the set of all TTSP term graphs whose variable labels are mutually distinct. For a TTSP term graph g, the TTSP graph language of g, denoted by L(g), is the set of all TTSP graphs obtained from g by substituting arbitrary TTSP graphs for all variables in g. Firstly, when a TTSP graph G and a TTSP term graph g are given as inputs, we present a polynomial time matching algorithm which decides whether or not L(g) contains G. The minimal language problem for the class

Polynomial time inductive inference of ordered tree languages with height-constrained variables from positive data

\mathcal{L_{TTSP}}