This work aims at helping students understand complex mathematical expressions that appear in academic documents such as research papers. To this end, we focus on subexpressions of an expression given as a query, and try to annotate them using information about known similar expressions, while most existing approaches tend to find out an expression similar to the whole of a query. Assuming that an expression is given in the form of Content Markup of MathML, we construct its DOM tree, and extract subtrees each of which corresponds to one of its subexpressions from the whole tree. Then, we search a database storing known expressions associated with their meta-data for ones similar to the subexpressions. To evaluate the usefulness of our approach, we conducted experiments using actual mathematical expressions collected from a web site and a mathematical textbook, and confirmed that considering subexpressions can bring on more information helpful for better understanding of a given query expression compared to annotating itself.

JADER database collects spontaneous reports of adverse effects of drugs in Japan. We analyze this data using the ATC classification hierarchy of drugs developed by WHO as the explanatory attribute. We propose a new method to extract characteristic nodes out of this hierarchy responsible for a specific adverse effect. PharmCompo database has been developed which enables the accurate counting of drug names appearing in various drug databases. ATC codes were also attached to all entries of PharmCompo. Application to anaphylaxis reaction has extracted 9 classification nodes including serotonin antagonists, first-generation cephalosporins and local anesthetics. Eighteen drugs were also indicated to lead to high anaphylaxis ratios, while their brother drugs in the same classification show low ratios. The comparison of chemical structures in the same class suggested that an amide-A-A-N fragment tends to cause the adverse effect.

Information diffusion over a social network can be modeled as stochastic processes of state changes. In this paper, we propose an information diffusion model that takes into account topics of information. More specifically, the proposed model determines the diffusion probability for a directed link by using the content attribute of a target document that will propagate over the link, which represents the topic distribution in the document, and the link attribute that expresses the topic distribution in documents that have propagated over the link. The number of model parameters to be learned is only twice of the number of topics considered, which is much less than the one for traditional models, and they can be efficiently and accurately learned from observed diffusion sequences based on the framework of the maximum likelihood estimation. Through an experiment using real world retweet sequences, we confirmed that the proposed model allows us to estimate the length of an information diffusion sequence more accurately compared to an existing model that do not consider topics at all.

In this paper, we tackle a problem for clustering a graph sequence to discover changing clusters. In the conventional method, spectral clustering with k-means is applied to this problem. In this paper, we replace the k-means to fuzzy c-means to detect the degree to which each vertex in a graph belongs to clusters in the situations of changing clusters. The experimental result shows that our method detects clusters merging with time from a graph sequence.

In this paper, we propose an efficient method for indexing and searching graphs for the supergraph component search problem. The problem is to find graphs contained by a given query graph from a graph database consisting of many graphs. While the conventional methods use frequent subgraph patterns in the database to construct an index for the database, the proposed method uses subgraph patterns which are infrequent for its index. In our experimental evaluation, we show one order of magnitude improvement by our proposed method compared with the conventional method.