Hiroki Sakaji

Text-Visualizing Neural Network Model: Understanding Online Financial Textual Data

This study aims to visualize financial documents to swiftly obtain market sentiment information from these documents and determine the reason for which sentiment decisions are made. This type of visualization is considered helpful for nonexperts to easily understand technical documents such as financial reports. To achieve this, we propose a novel interpretable neural network (NN) architecture called gradient interpretable NN (GINN). GINN can visualize both the market sentiment score from a whole financial document and the sentiment gradient scores in concept units. We experimentally demonstrate the validity of text visualization produced by GINN using a real textual dataset.

A Topic Trend on P2P Based Social Media

This paper shows a topic trend on a P2P based Social Network Service. There is a text-based Social Network Service (SNS) named Mastodon. Mastodon is a peer-to-peer and open-source SNS. Many persons and companies run Mastodon instances. We consider that there is a topic trend for each node. In this paper, we collect text messages and infer topic trend on a Mastodon instance using Latent Dirichlet Allocation(LDA). The understanding a topic trend helps to choice an instance that a user should participate.

A Web-Based Visualization System for Interdisciplinary Research Using Japanese Local Political Corpus

This paper describes a web-based visualization system, for interdisciplinary research, using the Japanese local political corpus. We illustrate the system for the corpus, which contains the local assembly minutes of 47 Japanese prefectures from April 2011 to March 2015. This four-year period coincides with the office term for assembly members in most local governments. Our system provides full-text search features for utterances, context word extraction using Key Words in Context (KWIC), map visualization, cross-tabulation tables, and political keyword extraction using TF–IDF. We endowed the system with these features to promote its wide range use in various research fields.

Extracting Laboratory Front Pages from University Websites

In this paper, we propose a method for extracting laboratory front pages from university websites. There are more than 779 universities and colleges in Japan. For selecting a university or a college, some high school students want to know what laboratories these universities or colleges have. To learn about these laboratories, high school students have to search the laboratory front pages from the university websites. However, sometimes it is difficult to find a laboratory front page because they are sometimes buried deep in the hierarchy of university websites. Our method extracts laboratory front pages by using a support vector machine model and applying certain rules. We also developed a laboratory search system that can be used to retrieve laboratory front pages extracted with our method. We evaluated our method and confirmed that is attained 85.0% precision and 65.5% recall.

A Method for Extracting Correct Links from Automatic Created Links on Folksonomy

A folksonomy is a system to classify an online item/concept by a short text label as a tag that is applied by users. Many folksonomies include many tags that consist words of long-tail contents. To explain these tags, A wiki-like system provided by some folksonomy services for user understanding. These systems provide a function that creates a link to another article automatically that described in an article for enriching the articles. The cause of the automatic link construction function, those wiki-like systems includes many incorrect links. By these incorrect links, it is hard to use the systems for applications that use Wikipedia internal links. Therefore, we propose a method for extracting correct links from all of the automatically created links.