Takeru Yokoi

Emoticon Extraction Method Based on Eye Characters and Symmetric String

Emoticon is often used in short messages to shortly describe actions, feelings and so on. That can also represent sentimental intention such that its difficult to describe by only language. Recently the sentiment analysis has been focused in cases of election, economic market and so on. Consideration of emoticon is also useful in such cases and in the first place, emoticon extraction from text is the important step to analysis them. However, to extract the emoticon correctly is difficult since recent emoticon uses Unicode characters as compared to former emoticons characters. Therefore, we have proposed a novel approach to extract emoticon focusing on characters describing eyes and the symmetry of emoticons string.

Topic Extraction for a Large Document Set with the Topic Integration

We propose here a method to extract topics from a large document set with topic integration from some small document sets. In order to extract topics, the Non-negative Matrix Factorization (NMF) is applied to document sets.

Index Words Selection with ICA

We propose here a method to select index words for the construction of a document vector from a corpus using the independent component analysis (ICA). It is useful to select index words of a document vector since its dimension is large. The ICA is one of the methods in analyzing the latent semantics of documents. It is reported the independent components obtained by the ICA represent the topics in the documents. The words in the independent component are considered to be the key words of the topic. The proposed method selects the key words which have high weight in each independent component and adds them to a set of index words. In addition, we selected other words related to the key words according to the chi-squared measure between the co-occurrence of the key words and each word and the appearance of the key words, and have also added them to the set of index words. Finally, an evaluation of the index words obtained has been carried out.

Information recommendation using ICA

We describe an information filtering system using independent component analysis (ICA). A document–word matrix is generally sparse and has an ambiguity of synonyms. To solve this problem, we propose a method to use document vectors represented by independent components. An independent component generated by ICA is considered as a topic. In practice, we map the document vectors into a topics space. Since some independent components are useless for recommendation, we select the necessary components from all independent components by a maximum distance algorithm (MDA). Although Euclidean distance is usually used by MDA, we propose topic selection by cosine-distance-based MDA to solve the mismatch of similarities in information filtering. We create a user profile from the transformed data with a genetic algorithm (GA). Finally, we recommend documents with the user profile and evaluate the accuracy by imputation precision. We have carried out an evaluation experiment to confirm the practicality of the proposed method.

Information filtering using SVD and ICA

We propose an information filtering system for documents by a user profile using latent semantics obtained by singular value decomposition (SVD) and independent component analysis (ICA). In information filtering systems, it is useful to analyze the latent semantics of documents. ICA is one method to analyze the latent semantics. We assume that topics are independent of each other. Hence, when ICA is applied to documents, we obtain the topics included in the documents. By using SVD remove noises before applying ICA, we can improve the accuracy of topic extraction. By representation of the documents with those topics, we improve recommendations by the user profile. In addition, we construct a user profile with a genetic algorithm (GA) and evaluate it by 11-point average precision. We carried out an experiment using a test collection to confirm the advantages of the proposed method.