Multi-label classification in document data is the task of correctly assigning multiple class labels to each document. However, there is often a semantic hierarchical structure among the assigned labels, and considering the hierarchical structure can improve the accuracy of label prediction. The Multi-label Box Model (MBM) has been proposed as a multi-label classification model that takes into account the semantic hierarchical structure among labels, and its effectiveness has been demonstrated when class labels of all layers are assigned to training data. However, real-world document data posted on user-contributed websites often lack class labels for all layers of the hierarchy. If such data is used to train MBM, the accuracy of label prediction is reduced. In this study, we propose a framework for learning MBM after complementing labels of missing hierarchies by introducing Bidirectional Encoder Representations from Transformers (BERT). The effectiveness of the proposed method is also demonstrated through evaluation experiments, which compare the accuracy of the conventional method and the proposed method when applied to data with missing labels of some hierarchies.

In recent years, more and more used smartphones have been bought and sold through online sales services in the used smartphone market, and it is desirable to utilize the large amount of transaction data accumulated in conjunction with these transactions when listing and purchasing used smartphones. Used equipment buyers can use this data to analyze market price trends and the factors that determine those prices, which can lead to optimal purchase strategies and pricing. However, used smartphones are handled by various sales services. For such a target problem, it would be possible to understand which factors contribute to the selling price if a prediction model could be constructed to explain the selling price based on various characteristics. In this study, we analyze price determinants using a model that incorporates the gradient boosting method, which is a model with high accuracy and interpretability, with the help of explainable AI. In this analysis, it is undesirable to apply a single pricing factor analysis model that could not consider differences in sales services, which has been the focus of previous analyses of the used equipment market. Therefore, we proposes an analytical model that employs the technique of explainable AI for the different price determinants among sales services. The proposed model is applied to analyze actual sales data of used smartphones and discuss the results.

In general, Machine Learning does not ensure the proper prediction if the statistical structures of the features between training data and prediction data are different, but it is sometimes required to apply a prediction model to a target which may have the different structure from its train data. In recent years, the studies to address this challenge have been actively conducted, and one of them is Adversarial Discriminative Domain Adaptation(ADDA), which is the classification model with adversarial training of Generative Adversarial Networks(GAN). ADDA has a defect that it uses all data from mini-batch, which includes bad data for training. In this study, we propose the improved method of ADDA with the application of GAN's related study, Top-k training. This application would enable ADDA to select useful data based on internal outputs, and the prediction accuracy is expected to increase. The result of the experiment shows that proposed method has significances of the accuracy and the length of training time.

企業がマーケティング施策を講じる際、適切な効果検証を行い、適切な意思決定につなげることは非常に重要な課題であり、行動履歴データを活用した効果検証によるデータ駆動型マーケティングへの期待が高まっている。近年では、複数の種類の施策を同期間に実施する場合が多く、このような状況においても施策ごとの効果を正しく推定及び検証できる手法がマーケティング戦略立案のために求められている。そこで本研究では、複数種類の施策効果を同時に推定可能な手法を活用し、ファッションECサイト内のバナー広告の効果の推定及び分析を行う。

In recent years, marketing using attribute information associated with member accounts of online services has been widely used. However, the majority of users are non-members who use services without registering for an account, and it is difficult to implement measures using attribute information for these non-member users. In order to deal with this situation, semi-supervised learning is an effective way to increase the number of users with attribute information by predicting it from the history data of member users who have attribute information, using the history data of non-member users as well. One of such semi-supervised learning methods is the Ladder Network, which is a neural network based model with adding and removing noise. This model provides highly accurate prediction for image data, and is also considered to be useful for predicting user attributes from historical data, where the feature vector is high-dimensional. However, this method cannot be applied to the case where the label takes ordered value, such as the user's age category. In this study, we propose an extended model based on the Ladder Network that incorporates a mechanism that can appropriately predict the user's attribute information. We also conduct an evaluation experiment using actual browsing history data to show the effectiveness of the proposed method.

It is important for companies to verify the effects of their marketing measures and to make right decisions. To verify the effects from observational data correctly, they make use of causal inference. In recent causal inference, after allocating subjects to two groups and treating them differently, they often seek to estimate the Conditional Average Treatment Effect (CATE) to better understand causal mechanisms. CATE makes it possible to identify the group of users for whom it is effective to take measures. As a CATE estimation method, Causal Tree which has high interpretability and usefulness for analyzing the factors that affect measures, has been proposed. However, this method cannot be used when they allocate subjects to two groups on purpose due to selection bias. Therefore, we propose a method for estimating CATE taking account of selection bias based on Causal Tree. Finally, we evaluate the precision of CATE estimates by simulation experiments. In addition, we apply the proposed method to an actual data and show the usefulness of the proposed method.

Due to the accumulation of browsing history data on EC sites, Web marketing techniques are of growing significance. Most previous studies analyzed differences in overall browsing pages between purchasing and non-purchasing sessions by constructing a discriminative model and proposed measures for all users. However, it is difficult to utilize this model when considering personalized measures for each user. In this situation, a generative model, which infers browsing-behavior conditioned by whether a user purchases or not, is effective. Conditional VAE infers the data from the label and features of input data. In this paper, we apply Conditional VAE to browsing history data and identify important pages by generating a pseudo session assuming that a user in a non-purchasing session purchases. We propose a method to analyze important browsing pages that contribute to each user's purchase. We clarify the effectiveness of our proposed method by using real browsing history data.

We construct a decision support model that can help taxi drivers dispatch their vehicles appropriately based on machine learning by utilizing probe data of taxis in Tokyo. Traditionally, taxi dispatch has relied on the driver's experience and intuition. The number of customers depends on the knowledge gained through many years of experience. For examples, many taxis are sometimes waiting for customers, and sometimes many customers wait in a long queue for a taxi. In addition, there are differences in the transport distance depending on the location. However, in a given situation, not all drivers know places with high demand. Therefore, it is desirable to build a model, which is easy to understand for drivers, that enables efficient acquisition of customers regardless of their experience. In this situation, we propose an analysis model that supports driver's decision based on taxi probe data.

Recent years, semi-supervised learning which classifies test data into correct category using not only training (labeled) data but a large number of unlabeled data has paid attention. However, in the semi-supervised learning setting, there is a problem that classification accuracy deteriorates because distribution of labeled data is biased. The SemiBoost is one of semi-supervised learning method to solve the problem. The SemiBoost is a binary classification method. However, this method can not be extended directly to multi-class classification. In this research, we propose the way to extend the SemiBoost for multi-class classification using the concept of Error Correcting Output Code (ECOC) method. To verify the effectiveness of our proposed method, we conduct simulation experiment using UCI machine learning repository.