Updated on 2025/01/04

写真a

 
YAMAGIWA, Ayako
 
Affiliation
Faculty of Science and Engineering, School of Creative Science and Engineering
Job title
Research Associate

Awards

  • Best Paper Award, APIEMS 2022

    2022.11   The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022)   A New Analytical Model for Product Reviews Based on BERT Feature Extraction

    Winner： Koutarou Yamashita, Ayako Yamagiwa, Kyosuke Hasumoto, Masayuki Goto

 

Papers

  • An Improved Search Algorithm Based on Safe Exploration for Optimization to Control Worsening of Objective Function Value

    Yuka Nakamura, Taiga Yoshikawa, Ayako Yamagiwa, Masayuki Goto

    Industrial Engineering and Management Systems   23 ( 2 ) 125 - 135  2024.07  [Refereed]

    DOI

    Scopus

  • An Estimation Model of Intrinsic Evaluation Ratings by Customer Reviews Based on BERT Feature Extraction

    Kotaro Yamashita, Ayako Yamagiwa, Kyosuke Hasumoto, Masayuki Goto

    Industrial Engineering and Management Systems   23 ( 2 ) 182 - 194  2024.07  [Refereed]

    DOI

    Scopus

  • Motion Discrimination Model from Sensing Point Cloud Data based on One Dimensional CNN and Data Augmentation

    SUZUKI Kazuchika, MIZUTANI Miho, YAMADA Koki, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI   JSAI2024   2D1GS204 - 2D1GS204  2024

    　View Summary

    In recent years, the importance of sensing technology for the safety of residents in elderly care facilities has increased, and in particular, there are growing expectations for the use of time-series point cloud data acquired by millimetre-wave radar. Many methods based on deep learning have been proposed for motion discrimination using point cloud data. Here it is necessary to construct a feature space that contains sufficient information for identification and to prepare a large amount of supervised data. However, the number of points in the point cloud data acquired from millimetre-wave radar, which is the subject of research, varies with each acquisition time. Because of that, these point clouds cannot be used as input to the deep learning model as they are. In addition, there is the problem that the acquisition of data with motion label is costly. In this study, we propose an action recognition method that constructs a feature space using various statistical information over time and models it with a One Dimensional CNN that performs convolution in the time direction, along with preparing sufficient training data through data augmentation method to improve recognition accuracy. Furthermore, we apply the proposed method to actual point cloud data and confirm that the proposed method is capable of motion identification and that the accuracy is improved by data expansion.

    DOI

  • User Review Analysis Method from Word Perspective Based on Topic Model and Quality Element Classification

    OGASAWARA Noriko, FUCHI Ayuno, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI   JSAI2024   2B1GS201 - 2B1GS201  2024

    　View Summary

    In recent years, many methods have been studied to analyze large amounts of review data posted by users in order to improve products and services. Most of these methods are mainly studies that attempt to extract some kind of information from user reviews using various machine learning techniques, and most of them do not provide an analysis method that clearly prioritizes quality improvement by considering the characteristics of each factor and the customer's viewpoint. In contrast, studies focusing on quality improvement have proposed a review analysis that introduces the concept of the Kano model, which classifies quality factors into "attractive quality" and other categories based on changes in customer satisfaction depending on their presence or absence. However, these studies have the problem that they cannot analyze at the word level or focus on only some words. In this study, we propose a method to analyze all words in a review sentence and classify quality factors. Specifically, by combining LDA learned using review sentences and logistic regression, we quantify the influence of each word that appears in the review sentences on the evaluation value. Through analysis of real data, we show that the proposed method can be used for detailed word-level analysis and heuristic quality factor extraction.

    DOI

  • A Customer Analysis Method Considering the Diversity of Purchasing Behavior Based on the Knowledge Graph Attention Network

    FUJIWARA Daiki, MORIKAWA Takuya, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI   JSAI2024   2B1GS203 - 2B1GS203  2024

    　View Summary

    In recent years, e-commerce sites have been used to conduct customer analysis, with the aim of building good relationships with customers and improving long-term sales. Evaluating the diversity of customers' purchasing behavior is particularly important as a marketing angle. The scalar values assigned to each customer in studies analyzing purchasing behavior diversity do not consider the varying impact of individual purchase items on the indices. Therefore, customers who should be treated with different business measures may be treated as the same. This study proposes a method for customer analysis that considers the impact of each purchase item on the index. The method calculates features that represent the diversity of each customer's purchase behavior by utilizing the distribution of weights assigned to each purchase in the Knowledge Graph Attention Network, a type of recommendation model. The effectiveness of the proposed method is demonstrated by applying it to real data.

    DOI

  • A method for improving the accuracy of multi-domain adaptive vision language model using prompt learning

    GAO ZHENYU, YAMAGIWA AYAKO, GOTO MASAYUKI

    Proceedings of the Annual Conference of JSAI   JSAI2024   1B3GS205 - 1B3GS205  2024

    　View Summary

    Methods for analyzing image data associated with linguistic information have garnered recent attention but encounter challenges due to varying data quantities across different image domains. In response, LADS was proposed, a model trainable without relying on image data from domains with limited samples, utilizing the embedding space between images and text in image language models. While LADS often employs simple domain description text, adequate text can improve model performance. To tackle this issue, we introduce CoOp, a method that optimizes the domain text in CLIP to enhance accuracy. CoOp achieves this by learning prompts, improving vision language models, and elevating CLIP accuracy. We expect the resulting prompts to represent diverse domains within LADS effectively. Finally, we validate the efficacy of our proposed method by applying it to actual data, demonstrating its ability to address imbalanced data quantities across various image domains.

    DOI

  • Analysis of Subjective Differences among Raters in the Estimation of Product Image Values Based on Pairwise Comparison Deep Neural Network

    YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI   JSAI2024   2B1GS204 - 2B1GS204  2024

    　View Summary

    In creating product maps, which are effective as a method of product analysis, the authors have shown that using a pairwise comparison DNN to estimate human evaluations significantly improves the efficiency of map creation relative to the number of data acquisitions compared to conventional creation methods. Specifically, assuming that the ratio of the evaluation values of two product image data to be compared can be obtained in a one-to-one comparison, a deep learning model can be trained using a realistic number of comparison results, and the estimated values can be used to estimate the evaluation values for an arbitrary evaluation axis for all product images. Here, subjective axes such as "cute" and "gorgeous" can be considered appropriate for the product map to be analyzed. When such personal axes are used, differences due to the subjectivity of the evaluators are expected to occur in the product image evaluation values. Therefore, in this study, we analyze how the subjectivity of the evaluators affects the estimation results of the DNN model by using a pair-wise comparison between multiple evaluators. Furthermore, we show that it is possible to analyze the similarity of evaluators' preferences by using the parameters of the deep learning model for pair-wise comparison.

    DOI

  • A Model for Analyzing Changes in Child-rearing Life Stage Concerns Based on Question Data in Dedicated Applications

    Koki Yamada, Yosuke Takao, Ayako Yamagiwa, Masayuki Goto

    Journal of Japan Industrial Management Association   74 ( 4 ) 153 - 166  2024.01  [Refereed]

    　View Summary

    Dedicated applications (apps) for child-rearing question and answer (QA) system have become common tools widely used by many parents in Japan. A user can post his/her various questions about child-rearing on the systems and exchange information with other parents online. Since the value of such dedicated applications increases through being used by a large number of users, it is considered important to continuously provide attractive functions to increase user satisfaction. Here, concerns of parents should be changed depending on their children's life stages. In addition, it can be assumed that users always have concerns of child-rearing even when they do not post any questions on the application. Therefore, analysis and comprehension of the problems of parents rearing small children by taking into account these properties will lead to the improving the users' satisfaction of QA services. For instance, the app can provide users with proper information with appropriate timing based on the analysis result. Although topic models are widely known to analyze the content of text data, this is not a sufficient method for visualizing changes in topics over time from question text data, which are relatively short sentences posted on dedicated apps for child-rearing QA systems. In this study, the authors propose an analytical model that can extract the topics of users' concerns from question data and compile topic transitions throughout the various stages of child-rearing. Furthermore, by taking a probabilistic view of the topic transitions, a method for estimating the topic that is most likely to shift from a previous topic at a given time to the next non-question period is constructed. Finally, the authors show the results of applying the proposed method to an actual data set and clarify the usefulness of the estimation result using the proposed method. The proposed method makes it possible to analyze how the worries and concerns of parents raising children evolve over time, and to gain various new insights into the perspectives of support for these users through apps.

    DOI

    Scopus

    1
    Citation
    (Scopus)

  • Active Learning Method for Pairwise Comparison Data

    Ayako Yamagiwa, Masayuki Goto

    5th IEEE International Conference on Cybernetics, Cognition and Machine Learning Applications, ICCCMLA 2023     310 - 315  2023

    　View Summary

    Companies need to create a product position map and correctly grasp the current situation. In the recent increase in the number of e-commerce sites, the influence of product images on customers' willingness to purchase is huge. In other words, if we can construct a product map based on customers' impressions of product images, we can expect to make the product lineup on e-commerce sites more efficient and attractive. Semantic Differential (SD) method and Multi-dimensional scaling (MDS) have been conventionally used to create product maps. However, these methods are unsuitable for analyzing a target with a large number of products, such as EC sites, because the number of product evaluation data required for the analysis increases depending on the number of products. Therefore, the authors proposed a method to construct a product image map of product images with high accuracy from a relatively small number of pairwise comparison data. Specifically, a model that estimates the pairwise comparison data by subjects is machine-learned. The learned model is used to supplement the missing pairwise comparison data to estimate the evaluation values of product images for a specific axis. The values are used to construct a map. The method of selecting a small number of data used to train the model can quickly identify the overall trend. Therefore, this study proposes an adequate data selection criterion to improve the accuracy of the map based on the concept of active learning. Experiments show that, in addition to uncertainty considerations, it is adequate to have a uniform amount of pairwise comparison data for each subject in the problem of complementing pairwise comparison data between subjects, as in the case of this study.

    DOI

    Scopus

  • An Analytical Model of Response Interval Between Employees on Business Chat Systems Based on Latent Class Model

    Fuyu Saito, Ayako Yamagiwa, Tianxiang Yang, Masayuki Goto

    Total Quality Science   7 ( 3 ) 149 - 160  2022.05  [Refereed]

    DOI

  • An Analytical Model Based on Purchase History for Products with Low Multiple Purchases from Each Customer

    YAMAGIWA Ayako, KUMOI Gendo, GOTO Masayuki

      J105-D ( 5 ) 297 - 309  2022.05  [Refereed]

    　View Summary

    A lot of study analyses product or customer behavior based on purchase history. The models based on co-occurrence of products and customer are one of most popular fields. However, analysis about product lucking of co-bought volume by same customer is difficult to be applied such models. In this study, we consider the combinations of products, customer's purposes, and customer's features are related to customer's preference. Then we propose the analytical model for such kinds of data with Factorization Machine classifying data whether buy the product or not. Finally, analysis result about flower product data shows the effectiveness of our proposal.

    DOI

  • Evaluation of Analysis Model for Products with Coefficients of Binary Classifiers and Consideration of Way to Improve

    Ayako Yamagiwa, Masayuki Goto

    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)   13316 LNCS   388 - 402  2022

    　View Summary

    Purchasing actions on e-commerce sites have become very common for general consumers in recent years. Products that were used to be bought at offline shops are purchased are also handled. Such products, like gifts or durable consumer goods, are often purchased infrequently and whose prefer items change each time they are purchased. A lot of methods are proposed for analysis purchase history data in order to improve customer satisfaction. However, most of them focus on the co-occurrence relationship between customers and products and treat products purchased by the same customer as similar. Then, it is difficult to use the conventional product analysis methods that have been proposed for purchase history data is difficult for some kinds of data mentioned before. Therefore, the authors have proposed an analysis method with extracting features of products by using the coefficients of binary classifiers that discriminates product purchases or not. In this study, we conduct experiments with artificial data in order to evaluate our method. Specifically, we verify how accurately the coefficients can be estimated and under what circumstances they can be estimated more accurately.

    DOI

    Scopus

▼display all

Presentations

  • Motion Discrimination Model from Sensing Point Cloud Data based on One Dimensional CNN and Data Augmentation

    SUZUKI Kazuchika, MIZUTANI Miho, YAMADA Koki, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2024

    Event date：
    2024
    　
    　

    　View Summary

    In recent years, the importance of sensing technology for the safety of residents in elderly care facilities has increased, and in particular, there are growing expectations for the use of time-series point cloud data acquired by millimetre-wave radar. Many methods based on deep learning have been proposed for motion discrimination using point cloud data. Here it is necessary to construct a feature space that contains sufficient information for identification and to prepare a large amount of supervised data. However, the number of points in the point cloud data acquired from millimetre-wave radar, which is the subject of research, varies with each acquisition time. Because of that, these point clouds cannot be used as input to the deep learning model as they are. In addition, there is the problem that the acquisition of data with motion label is costly. In this study, we propose an action recognition method that constructs a feature space using various statistical information over time and models it with a One Dimensional CNN that performs convolution in the time direction, along with preparing sufficient training data through data augmentation method to improve recognition accuracy. Furthermore, we apply the proposed method to actual point cloud data and confirm that the proposed method is capable of motion identification and that the accuracy is improved by data expansion.

  • User Review Analysis Method from Word Perspective Based on Topic Model and Quality Element Classification

    OGASAWARA Noriko, FUCHI Ayuno, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2024

    Event date：
    2024
    　
    　

    　View Summary

    In recent years, many methods have been studied to analyze large amounts of review data posted by users in order to improve products and services. Most of these methods are mainly studies that attempt to extract some kind of information from user reviews using various machine learning techniques, and most of them do not provide an analysis method that clearly prioritizes quality improvement by considering the characteristics of each factor and the customer's viewpoint. In contrast, studies focusing on quality improvement have proposed a review analysis that introduces the concept of the Kano model, which classifies quality factors into "attractive quality" and other categories based on changes in customer satisfaction depending on their presence or absence. However, these studies have the problem that they cannot analyze at the word level or focus on only some words. In this study, we propose a method to analyze all words in a review sentence and classify quality factors. Specifically, by combining LDA learned using review sentences and logistic regression, we quantify the influence of each word that appears in the review sentences on the evaluation value. Through analysis of real data, we show that the proposed method can be used for detailed word-level analysis and heuristic quality factor extraction.

  • A Customer Analysis Method Considering the Diversity of Purchasing Behavior Based on the Knowledge Graph Attention Network

    FUJIWARA Daiki, MORIKAWA Takuya, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2024

    Event date：
    2024
    　
    　

    　View Summary

    In recent years, e-commerce sites have been used to conduct customer analysis, with the aim of building good relationships with customers and improving long-term sales. Evaluating the diversity of customers' purchasing behavior is particularly important as a marketing angle. The scalar values assigned to each customer in studies analyzing purchasing behavior diversity do not consider the varying impact of individual purchase items on the indices. Therefore, customers who should be treated with different business measures may be treated as the same. This study proposes a method for customer analysis that considers the impact of each purchase item on the index. The method calculates features that represent the diversity of each customer's purchase behavior by utilizing the distribution of weights assigned to each purchase in the Knowledge Graph Attention Network, a type of recommendation model. The effectiveness of the proposed method is demonstrated by applying it to real data.

  • A method for improving the accuracy of multi-domain adaptive vision language model using prompt learning

    GAO ZHENYU, YAMAGIWA AYAKO, GOTO MASAYUKI

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2024

    Event date：
    2024
    　
    　

    　View Summary

    Methods for analyzing image data associated with linguistic information have garnered recent attention but encounter challenges due to varying data quantities across different image domains. In response, LADS was proposed, a model trainable without relying on image data from domains with limited samples, utilizing the embedding space between images and text in image language models. While LADS often employs simple domain description text, adequate text can improve model performance. To tackle this issue, we introduce CoOp, a method that optimizes the domain text in CLIP to enhance accuracy. CoOp achieves this by learning prompts, improving vision language models, and elevating CLIP accuracy. We expect the resulting prompts to represent diverse domains within LADS effectively. Finally, we validate the efficacy of our proposed method by applying it to actual data, demonstrating its ability to address imbalanced data quantities across various image domains.

  • Analysis of Subjective Differences among Raters in the Estimation of Product Image Values Based on Pairwise Comparison Deep Neural Network

    YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2024

    Event date：
    2024
    　
    　

    　View Summary

    In creating product maps, which are effective as a method of product analysis, the authors have shown that using a pairwise comparison DNN to estimate human evaluations significantly improves the efficiency of map creation relative to the number of data acquisitions compared to conventional creation methods. Specifically, assuming that the ratio of the evaluation values of two product image data to be compared can be obtained in a one-to-one comparison, a deep learning model can be trained using a realistic number of comparison results, and the estimated values can be used to estimate the evaluation values for an arbitrary evaluation axis for all product images. Here, subjective axes such as "cute" and "gorgeous" can be considered appropriate for the product map to be analyzed. When such personal axes are used, differences due to the subjectivity of the evaluators are expected to occur in the product image evaluation values. Therefore, in this study, we analyze how the subjectivity of the evaluators affects the estimation results of the DNN model by using a pair-wise comparison between multiple evaluators. Furthermore, we show that it is possible to analyze the similarity of evaluators' preferences by using the parameters of the deep learning model for pair-wise comparison.

  • A Clustering Method Using Embedded Representations Based on User Ratings

    Miho Mizutani, Ayako Yamagiwa, Masayuki Goto

    The 7th Asian Conference of Management Science and Application (ACMSA 2023) 

    Event date：
    2023.12
    　
    　

  • Multi-Task Learning for Estimating Consumer Impressions of Product Images

    Ayako Yamagiwa, Masayuki Goto

    The 7th Asian Conference of Management Science and Application (ACMSA 2023) 

    Event date：
    2023.12
    　
    　

  • Online Flipped Conference Based Data Science Education Program and Its Educational Effectiveness in Multi-University Collaboration

    Masayuki Goto, Manabu Kobayashi, Takeshi Moriguchi, Yoichi Seki, Hideo Suzuki, Takashi Namatame, Kazuhide Nakata, Aya Ishigaki, Masao Ueda, Kimitoshi Sato, Kenta Mikawa, Haruka Yamashita, Tomoaki Tabata, Tianxiang Yang, Ayako Yamagiwa, Yutaka Tajiri

    The 7th Asian Conference of Management Science and Application (ACMSA 2023) 

    Event date：
    2023.12
    　
    　

  • 求人データに対する適切な職種情報付与のためのラベル修正アルゴリズム

    山田晃輝, 山極綾子, 後藤正幸

    第46回情報理論とその応用シンポジウム (SITA2023) 

    Event date：
    2023.11
    -
    2023.12

  • 機械学習に基づく中古スマートフォン端末の将来価格予測モデルに関する一考察

    増田雅樹, 山極綾子, 後藤正幸

    第46回情報理論とその応用シンポジウム (SITA2023) 

    Event date：
    2023.11
    -
    2023.12

  • 施策効果の高い顧客グループの特定を目的とした機械学習に基づく実験計画手法

    中村友香, 山極綾子, 佐々木北都, 後藤正幸

    第46回情報理論とその応用シンポジウム (SITA2023) 

    Event date：
    2023.11
    -
    2023.12

  • オンライン求人サイトにおける適切な職種情報付与のためのラベル修正手法

    山田晃輝, 山極綾子, 後藤正幸

    第5回日本イーコマース学会 全国研究発表大会 

    Event date：
    2023.11
    　
    　

  • 能動学習を用いた商品画像ランキング推定と画像マップ作成

    山極綾子, 後藤正幸

    第5回日本イーコマース学会 全国研究発表大会 

    Event date：
    2023.11
    　
    　

  • 評価値との関連に着目した商品レビュー分析のための感情語抽出手法

    小笠原 のりこ, 泓 亜由乃, 山極 綾子, 後藤 正幸

    日本計算機統計学会シンポジウム論文集  東京 : 日本計算機統計学会

    Presentation date： 2023.11

    Event date：
    2023.11
    　
    　

  • 明示的評価を用いた埋め込み表現によるクラスタリング手法

    水谷美穂, 山極綾子, 後藤正幸

    日本経営工学会 2023年秋季大会 

    Presentation date： 2023.10

  • A Predictive Model of Future Sales Prices for Second-hand Smartphone Devices

    Masaki Masuda, Ayako Yamagiwa, Masayuki Goto

    23rd Asia Pacific Industrial Engineering & Management System Conference (APIEMS 2023) 

    Event date：
    2023.10
    　
    　

  • An Efficient Active Learning Approach based on BERT for Job;Label Classification Problem

    Koki Yamada, Ayako Yamagiwa, Goto Masayuki

    23rd Asia Pacific Industrial Engineering & Management System Conference (APIEMS 2023) 

    Event date：
    2023.10
    　
    　

  • Experimental Design based on Machine Learning for Finding Customer Groups with High Measure Effects

    Yuka Nakamura, Ayako Yamagiwa, Hokuto Sasaki, Masayuki Goto

    23rd Asia Pacific Industrial Engineering & Management System Conference (APIEMS 2023) 

    Event date：
    2023.10
    　
    　

  • Selection Criteria of Evaluation Axes for Impression Map Visualizations of Product Images

    Ayako Yamagiwa, Masayuki Goto

    23rd Asia Pacific Industrial Engineering & Management System Conference (APIEMS 2023) 

    Event date：
    2023.10
    　
    　

  • A Proposal of Product Image Analysis Model based on CVAE Learning Abstract Information of Product Descriptions

    Ayuno Fuchi, Masaki Masuda, Masakazu Asano, Ayako Yamagiwa, Masayuki Goto

    The 21st Asian Network for Quality Congress (ANQ 2023) 

    Event date：
    2023.10
    　
    　

  • An Approach of Customer Segmentation with SHAP Values Focusing on Individual Feature Influences on Outcome Variable

    Naru Shimizu, Yuka Nakamura, Ayako Yamagiwa, Masayuki Goto

    The 21st Asian Network for Quality Congress (ANQ 2023) 

    Event date：
    2023.10
    　
    　

  • 段階的パフォーマンス向上を目的とした人事評価データ分析

    浅野正和, 山極綾子, 後藤正幸

    経営情報学会 2023年度年次大会 

    Presentation date： 2023.06

  • Proposal of Product Image Generation Model based on CVAE Learning Abstract Information of Text Data

    FUCHI Ayuno, MASUDA Masaki, ASANO Masakazu, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2023

    Event date：
    2023
    　
    　

    　View Summary

    Each product on the EC site has attributes, image, and description and especially visual information has a large impact on customers' purchase decisions. Therefore, several studies to generate product images have been done recently. However, the abstract needs of customers cannot be reproduced with the generated images based on alone the specific attributes. On the other hand, some words included in the descriptions may express the customers' abstract needs. Therefore, if we can model the impact of abstract information expressed in descriptions on images and generate product images based on abstract information, we can capture customers' needs better. In this study, we propose an image generation method that combines Latent Dirichlet Allocation and Conditional Variational Autoencoder. We show the results of applying the proposed method to the real data and point out that it enables to generate product images based on abstract information extracted from product descriptions.

  • Proposal of customer segmentation method based on the impact of each feature on outcome variable

    SHIMIZU Naru, NAKAMURA Yuka, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2023

    Event date：
    2023
    　
    　

    　View Summary

    Customer segmentation is important for implementing appropriate marketing strategies to meet the different needs of each customer group. The purpose of customer segmentation is to improve the effectiveness of marketing strategies by implementing appropriate measures for each segment, and the formation of similar segments is required to determine the factors that determine the effectiveness of the measures. However, conventional methods do not fully consider this. Therefore, in this study, we propose a method of clustering similar customers based on the impact of feature variables on the effectiveness of measures by using SHAP value vectors, which are known as interpretation methods for machine learning models. This allows us to consider the similarity of the factors that determine the effectiveness of measures, making it possible to implement the most effective measures for each customer segment. We conducted experiments using artificial and actual data to demonstrate the effectiveness of the proposed method.

  • 知識グラフと強化学習に基づく説明可能な推薦のための効率的な経路探索アルゴリズム

    楊 冠宇, 清水良太郎, 山極綾子, 後藤正幸

    第45回情報理論とその応用シンポジウム (SITA2022) 

    Event date：
    2022.11
    -
    2022.12

  • BERTの特徴量抽出に基づく製品レビュー分析モデルに関する一考察

    山下皓太郎, 山極綾子, 蓮本恭輔, 後藤正幸

    第45回情報理論とその応用シンポジウム (SITA2022) 

    Event date：
    2022.11
    -
    2022.12

  • トピック分布を活用した文脈付きバンディットアルゴリズムによる施策決定法

    松苗亮汰, 山極綾子, 後藤正幸

    第45回情報理論とその応用シンポジウム (SITA2022) 

    Event date：
    2022.11
    -
    2022.12

  • ミニバッチ学習によるスケーラブルな隠れセミマルコフモデルの推定手法に関する一考察

    高尾洋佑, 山極綾子, 山下 遥, 後藤正幸

    第45回情報理論とその応用シンポジウム (SITA2022) 

    Event date：
    2022.11
    -
    2022.12

  • Hawkes過程を用いた暗号資産の取引状況分析手法に関する一考察

    増田雅樹, 山極綾子, 後藤正幸

    日本経営工学会秋季大会 

    Event date：
    2022.11
    　
    　

  • トピック分布を用いた文脈付きバンディットアルゴリズム

    松苗亮汰, 山極綾子, 後藤正幸

    第4回 日本イーコマース学会 全国研究発表大会 

    Event date：
    2022.11
    　
    　

  • 子育てQA アプリのサービス向上のためのユーザの課題変化分析に関する一考察

    山田晃輝, 山極綾子, 高尾洋佑, 後藤正幸

    第4回 日本イーコマース学会 全国研究発表大会 

    Event date：
    2022.11
    　
    　

  • BERTの特徴量抽出に基づく製品レビュー分析モデル

    山下皓太郎, 山極綾子, 蓮本恭輔, 後藤正幸

    第4回 日本イーコマース学会 全国研究発表大会 

    Event date：
    2022.11
    　
    　

  • DNNを用いた画像ランキング学習手法

    山極綾子, 後藤正幸

    第4回 日本イーコマース学会 全国研究発表大会 

    Event date：
    2022.11
    　
    　

  • Efficient Effective Path Search Algorithm for Explainable Recommendation Based on Reinforcement Learning

    Guanyu Yang, Ryotaro Shimizu, Ayako Yamagiwa, Masayuki Goto

    The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022) 

    Event date：
    2022.11
    　
    　

  • A Method for Embedding Representations of Customer Feelings for Product Images

    Ayako Yamagiwa, Masayuki Goto

    The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022) 

    Event date：
    2022.11
    　
    　

  • A Method for Scalable Inference of Hidden semi-Markov Model by Batch Learning

    Yosuke Takao, Ayako Yamagiwa, Haruka Yamashita, Masayuki Goto

    The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022) 

    Event date：
    2022.11
    　
    　

  • Context Construction Based on Latent Dirichlet Allocation for Contextual Bandit Algorithm

    Ryota Matsunae, Ayako Yamagiwa, Masayuki Goto

    The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022) 

    Event date：
    2022.11
    　
    　

  • A New Analytical Model for Product Reviews Based on BERT Feature Extraction

    Koutarou Yamashita, Ayako Yamagiwa, Kyosuke Hasumoto, Masayuki Goto

    The 22nd Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS2022) 

    Event date：
    2022.11
    　
    　

  • SHAP値を用いたクラスタリングによる顧客セグメンテーション手法の提案

    清水 成, 中村 友香, 山極 綾子, 後藤 正幸

    日本計算機統計学会シンポジウム論文集  東京 : 日本計算機統計学会

    Presentation date： 2022.11

    Event date：
    2022.11
    　
    　

  • DNNを用いた画像ランキング学習の手法に関する一考察

    山極 綾子, 後藤 正幸

    日本計算機統計学会シンポジウム論文集  東京 : 日本計算機統計学会

    Presentation date： 2022.11

    Event date：
    2022.11
    　
    　

  • A Study on Search Problem Based on Bayesian Optimization to Control Worsening of Objective Function Value

    Yuka Nakamura, Taiga Yoshikawa, Ayako Yamagiwa, Masayuki Goto

    The 20th Asian Network for Quality Congress (ANQ2022), Beijing(Online) 

    Event date：
    2022.10
    　
    　

  • A Topic Transition Model of Parents Concerns over Child-rearing Stages based on Question Data on Dedicated App

    Koki Yamada, Yosuke Takao, Ayako Yamagiwa, Masayuki Goto

    The 20th Asian Network for Quality Congress (ANQ2022), Beijing(Online) 

    Event date：
    2022.10
    　
    　

  • Evaluation of Analysis Model for Products with Coefficients of Binary Classifiers and Consideration of Way to Improve

    Ayako Yamagiwa, Masayuki Goto

    Proceeding of 24th International Conference on Human-Computer Interaction (HCI International 2022), Online 

    Event date：
    2022.06
    -
    2022.07

  • ハウスホルダーフローを導入したEmbedded Topic Modelに関する一考察

    松苗, 亮汰, 山極, 綾子, 後藤, 正幸

    情報処理学会第84回全国大会講演論文集 

    Presentation date： 2022.02

    Event date：
    2022.02
    　
    　

    　View Summary

    トピックモデルは，単語をその共起関係からモデル化し，自動で文書のトピックを抽出可能な機械学習モデルである．Embedded Topic Model（ETM）はトピックモデルの一つであり，深層学習を用いてトピックと単語を同一空間上にEmbeddingする．ここで，文書ごとにトピック間に相関が存在すると考えられるが，ETMはトピック分布の推定時に各トピックが独立と仮定している．そこで本研究では，ETMのトピック割合に対してハウスホルダーフローを適用することで，トピック間の相関を表現可能にしたFlow-ETMを提案する．Flow-ETMは従来のETMよりも，より柔軟に入力文書に沿うトピック分布の生成が期待される．最後に文書データセットを用いた評価実験により，提案手法の有効性を確認した．

  • 購買アイテムを特定する分類器パラメータを用いた商品分析モデルに関する一考察

    山極, 綾子, 後藤, 正幸

    第84回全国大会講演論文集 

    Presentation date： 2022.02

    Event date：
    2022.02
    　
    　

    　View Summary

    近年ECサイト上の購買が増加し，購買履歴データから顧客の嗜好を分析し活用する企業が増えている．顧客が商品を購入する際には，嗜好に加え購入用途が影響する可能性がある．加えて，用途の影響が大きい贈答用などの商品では，その購入頻度が低いことがある．しかし従来の埋め込み手法は同一顧客が購入した商品を，その用途に関わらず類似した商品であると見なすため適用することが難しい．そういった商品群を対象とし，購買アイテムを特定する分類器パラメータを用いた商品分析モデルが提案されており，実データに対して一定の成果が得られている．本論文では，人工データを用いて上記手法が持つ特性を明らかにすることで，その有効性を示す．

  • An Analytical Model of Changes in Problems over Stages of Child-rearing from Question Data on a Dedicated Application

    YAMADA Koki, YAMAGIWA Ayako, TAKAO Yosuke, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2022

    Event date：
    2022
    　
    　

    　View Summary

    Dedicated Apps for child-caring QA system are widely used by many parents, on which users can ask questions about child-rearing. Problems that parents have should be changed depending on their children's life stage. If we can comprehend the problems that parents who grow small children have from the question data, it can contribute to the improvement of the users' satisfaction for a QA service by giving users proper information with correct timing based on problems they have. In this study, we propose an analytical model which can extract the topics of users' problems from question data and acquire their topic transitions over the stages of child-rearing. Furthermore, by taking a probabilistic view of the topic transitions, we construct a method for estimating the topic that is most likely to shift from a previous topic at a given time to the next non-question period. Finally, we show the results of applying the proposed method to the real data and conduct an evaluation experiment to show the usefulness of the estimation result by the proposed method.

  • A Study on Online Learning Based on Bayesian Optimization to Suppress Deterioration of Objective Function Value

    NAKAMURA Yuka, YOSHIKAWA Taiga, YAMAGIWA Ayako, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2022

    Event date：
    2022
    　
    　

    　View Summary

    In recent years, many recommender systems used in e-commerce sites estimate preferring items for each user based on the past log data, and show them as a list. The performance of these systems is evaluated by measuring whether presented recommendation lists match the preferences of customers. However, the recommendation is not implemented just once, but is in fact a continuous process. Thus, it is important to discuss the performance based on the cumulative loss for the entire recommendation series. Meanwhile, online learning is a framework that can handle such sequential recommendation and evaluation. However, the purpose of online learning is to improve the efficiency of learning in order to estimate condition of best values, and most methods do not consider the cumulative loss of recommendations. On the other hand, Safe Exploration for Optimization(SafeOpt) has recently been proposed as a method to perform exploration while suppressing the deterioration of the objective function. However, this method has a problem that the searchable range depends on the initial input. In this study, we extend the original SafeOpt by introducing GP-UCB, which is a method for global search, and propose a method to suppress the deterioration of the objective function for a wide search area. Finally, we show the effectiveness of the proposed method by generating artificial data and conducting experiments.

  • 複数の商品購買順序情報を考慮する拡張Translation-based Recommendationモデルの提案

    李 ア舒, 山極綾子, 楊 添翔, 後藤正幸

    第44回情報理論とその応用シンポジウム 

    Event date：
    2021.12
    　
    　

  • An Improved Translation-based Recommendation Considering Last Purchasing Sequence

    Yashu Li, Yamagiwa Ayako, Tianxiang Yang, Masayuki Goto

    2021 IEEE 12th International Workshop on Computational Intelligence and Applications (IEEE IWCIA2021), Online 

    Event date：
    2021.11
    　
    　

  • 隠れマルコフモデルに基づくユーザのWebサイト閲覧行動分析

    高尾 洋佑, 山極 綾子, 山下 遥, 後藤 正幸

    日本計算機統計学会シンポジウム論文集  東京 : 日本計算機統計学会

    Presentation date： 2021.11

    Event date：
    2021.11
    　
    　

  • Flow-ETM : トピック間の相関を表現したEmbedded Topic Model

    松苗 亮汰, 山極 綾子, 後藤 正幸

    日本計算機統計学会シンポジウム論文集  東京 : 日本計算機統計学会

    Presentation date： 2021.11

    Event date：
    2021.11
    　
    　

  • A Study on Ensemble Learning Model with Interpretability

    Taiga Yoshikawa, Ayako Yamagiwa, Tianxiang Yang, Masayuki Goto

    The 19th Asian Network for Quality Congress (ANQ2021) 

    Event date：
    2021.10
    　
    　

  • EC サイト上の購買行動における顧客嗜好変化の分析手法に関する一考察

    李 ア舒, 山極綾子, 楊 添翔, 後藤正幸

    日本経営工学会春季大会 

    Event date：
    2021.05
    　
    　

  • 解釈性を有するアンサンブル識別機の効率的な学習法に関する一考察

    良川太河, 山極綾子, 楊 添翔, 後藤正幸

    日本経営工学会春季大会 

    Event date：
    2021.05
    　
    　

  • 生花ECサイトを対象とした閲覧履歴に基づく購買行動分析に関する一考察

    楊 添翔, 鎌形祐志, 山極綾子, 後藤正幸

    日本経営工学会春季大会 

    Event date：
    2021.05
    　
    　

  • 潜在クラスモデルに基づくビジネスチャットアプリ上の従業員コミュニケーション分析

    齊藤芙佑, 山極綾子, 楊 添翔, 後藤正幸

    第19回日本データベース学会年次大会 DEIM2021 

    Event date：
    2021.03
    　
    　

  • 生花ECサイトの購買履歴に基づく商品特性分析モデル

    山極綾子, 雲居玄道, 後藤正幸

    第19回日本データベース学会年次大会 DEIM2021 

    Event date：
    2021.03
    　
    　

  • An Effective Constructive Algorithm of Single Decision Tree Preserving Predictive Performance of Ensemble Learning

    YOSHIKAWA Taiga, YAMAGIWA Ayako, YANG Tianxiang, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2021

    Event date：
    2021
    　
    　

    　View Summary

    Decision tree is an useful model for label classification and has high interpretability. However, the common size of training data prepared for a decision tree could lead to overfitting. Although the ensemble discriminator of decision tree prevents overfitting and earns high predictive accuracy, it will lose interpretability because of generating a large amount of random decision trees. Therefore, if we can learn a single decision tree that has the same predictive performance to the ensemble discriminator, it should be useful for actual application. In this study, we propose a method for learning a single decision tree with high accuracy with Autoencoder as a generative model, and we also use SMOTE as oversampling method to generate additional learning data by following the distribution of the target data with a small amount of computation. Finally, we show the effectiveness of the proposed method by actual data.

  • 低頻度購買商品を対象とした分散表現モデリングに関する一考察

    山極綾子, 楊 添翔, 後藤正幸

    日本経営工学会秋季大会 

    Event date：
    2020.10
    　
    　

  • An Analytical Model of Response Interval Between Employees on Business Chat Systems Based on Latent Class Model

    Fuyu Saito, Ayako Yamagiwa, Tianxiang Yang, Masayuki Goto

    18th Asian Network for Quality Congress (ANQ2020) 

    Event date：
    2020.10
    　
    　

  • Latent class model for analyzing response intervals between users on business chat applications

    SAITO Fuyu, YAMAGWA Ayako, YANG Tianxiang, GOTO Masayuki

    Proceedings of the Annual Conference of JSAI  The Japanese Society for Artificial Intelligence

    Presentation date： 2020

    Event date：
    2020
    　
    　

    　View Summary

    These days many companies try to use communication data on business chat systems for the management of human resources. Especially, the reply time between a talker and a receiver is affected by the relationship with the receiver desired to establish. Therefore, the analysis model about the reply time between two employees as the minimum unit with using a chat system should be effective. In this research, we propose a latent class model that quantitatively expresses the relationship between the talker, the receiver and the reply time and analyze the communication characteristics from the viewpoint of the reply time. We can interpret the combination of high-dimensional data and the difference between user behavior by introducing multiple latent variables that are probabilistically occurred behind the observed variables. Finally, we apply the model to the real data of a Japanese company and show the usefulness of the model.

  • An analytical model of users' communication on a chat system

    Ayako Yamagiwa, Yuto Seko, Tianxiang Yang, Masayuki Goto

    The 20th Asia Pacific Industrial Engineering And Management Systems (APIEMS 2019) 

    Event date：
    2019.12
    　
    　

  • ビジネスチャットアプリ上の会話履歴データを対象としたトピック分析モデル

    山極綾子, 世古裕都, 楊 添翔, 後藤正幸

    第42回情報理論とその応用シンポジウム(SITA2019) 

    Event date：
    2019.11
    　
    　

  • Multi-Valued Classification Based on ECOC with Support Vector Machine

    Ayako Yamagiwa, Haruka Yamashita, Masayuki Goto

    17th Asian Network for Quality Congress (ANQ2019) 

    Event date：
    2019.10
    　
    　

  • サポートベクトルに着目したECOC-SVMによる多値分類に関する一考察

    山極綾子, 馬賀嵩士, 山下 遥, 後藤正幸

    日本経営工学会春季大会 

    Presentation date： 2016.05

▼display all

 

Internal Special Research Projects

  • 深層学習モデルに基づく商品画像データイメージ分析手法に関する研究

    2023  

    　View Summary

    One of the most important situations in which data is used in corporate activities is in marketing activities, where it is important not only to analyze customer purchasing behavior but also to optimize the company's product lineup to meet the needs of consumers. Although many product position maps have been proposed in the past, all of them require human sensitivity data for each product, which is expensive to analyze, and there is no useful model that can evaluate multiple products efficiently and from a bird's eye view. Therefore, the objective of this research is to develop a low-cost technology to construct a deep learning model that evaluates the impression or image (emotional quality) that consumers have of product images, and to empirically demonstrate that this model can be a powerful tool for visualizing the emotional quality of multi-products and optimizing product lineups.In FY2023, we proposed a deep learning model for evaluating product images as the basis of our research, applied it to real data, and analyzed the results in a paper. Unlike conventional artificial intelligence approaches to images, which recognize objective concepts such as "dog" or "cat," this method is an artificial intelligence that evaluates images subjectively, such as "cute" or "vivid". While Kansei Engineering is a well-known analysis method for subjective sensitivity evaluation, it requires detailed analysis of all products and evaluation of combinations of products. Therefore, we proposed a method to evaluate the sensibility of all product images at a low cost using deep learning, and examined the method's effectiveness using actual product image data. We examined the effectiveness of the proposed method using actual product image data. The results showed that even when the number of evaluations is small, it is possible to create product position maps. In addition to the above research paper, we also studied improvement ideas to make the proposed method a more powerful tool, and presented the results both domestically and internationally. For example, regarding data selection for efficient model training, we showed that accuracy can be improved by obtaining ratings for as many different product images as possible, considering the fact that the data used are the results of pairwise comparisons between product images. We have also studied the results of combining multiple sensibility evaluation indices and confirmed that the accuracy of estimating the evaluation values of product images can be improved depending on the combination of indices.