The purpose of this study is to employ a novel mixed method to better understand the differences in the customer service experience of the digital banking services in South Korea and the Philippines. Data mining techniques and customer journey mapping analysis were utilized to understand the proposed issues. The results indicate that there are four critical significant points of digital banking services between South Korea and the Philippines including the number of touchpoints, speed of results, registration requirements, and touchpoint deviations. Potential causes and implications are discussed in this article. The contribution of this study is using mixed approach to understand the issues which related to bank marketing in the digital era. Additionally, this study also enriches the investigations of customer service experience in banking across different countries. Overall, the findings of this study benefit the development of digital banking services, especially in the Asia Pacific countries.

Due to the rapidly growth of social media, potential moviegoers always depend on the online reviews to make their purchase decisions. Film companies need to know what aspects of reviews will drive sales up or down. This study proposes a framework which integrates an aspect-based sentiment analysis and econometric modeling to explore the relationships between the information features in online reviews and movie ticket sales. The empirical results indicate that whereas the rating does not matter to moviegoers and does not affect movie revenues, additional textual review, both positive and negative contents, does have a positive impact on moviegoers, and further prompts the movie revenues. These findings have significant implications for movie producers as well as advertisers to target promotions at their audience accordingly.

High-fidelity kinship face synthesis receives increasing interest in this technology for visual kinship applications, including law enforcement, social media analysis, finding lost children, etc. However, it is a challenging task because of the unresolved ambiguities by the limited amount of available kinship data and severe data noise. To address these issues, we leverage the pretrained state-of-the-art face synthesis model, StyleGAN2, to assist the synthesis. With StyleGAN2, we develop three different kinship face synthesis strategies: (1) synthesis based on the kinship statistics, (2) synthesis using the latent code interpolation of the parents, and (3) synthesis based on the latent code interpolation from the disentangled age and gender independent latent space. The first two methods synthesize kinship faces through the direct manipulation of the original StyleGAN2 latent codes. The third one, on the other hand, is a two-stage synthesis method which first learns an age and gender invariant latent representation upon the one of StyleGAN2 to represent the genes. Combining with the maximal selection process to fuse the corresponding representation of parents, we form the genes of the child followed by them feeding back to StyleGAN2 for the final synthesis. With extensive ablation studies and experiments, we observe that all three methods can generate more photo-realistic and clearer faces than the previous state-of-the-art method. In addition, the third method achieves the best kinship verification results on the FIW dataset. Surprisingly, the subjective evaluation results of the three proposed methods are very close because typical humans are not good at recognizing unfamiliar kinship faces.

With the upcoming edge AI, federated learning (FL) is a privacy-preserving framework to meet the General Data Protection Regulation (GDPR). Unfortunately, FL is vulnerable to an up-to-date security threat, model poisoning attacks. By successfully replacing the global model with the targeted poisoned model, malicious end devices can trigger backdoor attacks and manipulate the whole learning process. The traditional researches under a homogeneous environment can ideally exclude the outliers with scarce side-effects on model performance. However, in privacy-preserving FL, each end device possibly owns a few data classes and different amounts of data, forming into a substantial heterogeneous environment where outliers could be malicious or benign. To achieve the system performance and robustness of FL's framework, we should not assertively remove any local model from the global model updating procedure. Therefore, in this paper, we propose a defending strategy called FedEqual to mitigate model poisoning attacks while preserving the learning task's performance without excluding any benign models. The results show that FedEqual outperforms other state-of-the-art baselines under different heterogeneous environments based on reproduced up-to-date model poisoning attacks.

With the popularity of AIoT (Artificial Intelligence of Things) services, we can foresee that smart end devices will generate tremendous user data at the edge. In particular, it is critical to address how to properly distill knowledge from the edge network in a communication-efficient and privacy-preserving manner. Federated learning (FL), one of the promising machine learning frameworks, ensures data privacy by allowing end devices to collaboratively train a shared model without exposing raw data to an aggregation server. However, due to its distributed nature, the framework is vulnerable to two major threats: the Model Inversion Attacks and the Model Poisoning Attacks. An abnormal aggregator or malicious end devices may probably launch these attacks in the training phase. The former leaks sensitive information by reversing the model weights to users' raw data. Still, the latter can break the model security and mislead the global model to wrong inference results. Unfortunately, the existing research has not tackled such two-sided model attacks that occurred concurrently in FL. Therefore, in this paper, we propose a dual-masking federated learning (DMFL) framework that advocates partial weights uploading in the aggregation process and applies two kinds of masks on both the end device and the aggregator sides. Based on the benchmark data for image classification, our experimental results show that the proposed DMFL framework outperforms other baselines, confirming that it can successfully preserve weights privacy and protect model security for AIoT.

Recently, the rapid growth in the number of customer reviews on e-commence platforms and in the amount of user-generated content has begun to have a profound impact on customer purchasing decisions. To counter the negative impact of social media marketing, some firms have begun hiring people to generate fake reviews which either promote their own products or damage their competitor's reputation. This study proposes a framework, which takes advantage of both supervised and unsupervised learning techniques, for the observation of behaviors among spammers. Then, based on the behavior of participants on web forums, the authors build up a post-reply network. The main focus is on the behavior-related features of the reviews, their propagation, and their popularity. The primary objective of this study is to build an effective online spammer detection model and the method detailed in this work can be used to improve the performance of spammer detection models. An experiment is carried out with a real dataset, the results of which indicate that these new features are important for identifying spammers. Finally, random walk clustering is applied to investigate the post-reply network. Some interesting and important features are observed in the interactions between a group of spammers which could be subjected to further research.

Digital banks are new entrants in the banking industry in the Philippines as they only started late 2018. Since then, a handful of players have and are still emerging. With more and more people becoming technologically savvy, it is very critical for financial institutions to develop a digital banking application that will stand out from the competition. This paper aims to use text mining methods to analyse digital banking application reviews. This study will perform topic modelling using LDA to explore customer concerns and will mine association rules between the digital banking features with the review score. The results will reveal which areas the digital banking application can further optimize for customer satisfaction and retention.

The spread of information on Facebook and Twitter is much more efficient than on traditional social media platforms. For word-of-mouth (WOM) marketing, social media have become a rich information source for companies or scholars to design models to examine this repository and mine useful insights for marketing strategies. However, social media language is relatively short and contains special words and symbols. Most natural language processing (NLP) methods focus on processing formal sentences and are not well-suited to such short messages. In this study we propose a novel sentiment analysis framework based on deep learning models to extract sentiment from social media. We collect data from which we compile a dataset. After processing these special terms, we seek to establish a semantic dataset for further research. The extracted information will be useful for many future applications. The experimental data have been obtained by crawling several social media platforms.

This article describes how the bank industry in Taiwan must function in today's tough and fiercely competitive domestic credit card market and subdued global market. Banks are increasingly emphasizing the importance of retaining customers in order to sustain market share and remain profitable. This study proposes a new model which local banks can use to detect potential customer churn and provide an early warning indicator of problems that could lead to loss of customers. The model incorporates a customer relationship management database with a built-in time factor and applied temporal abstraction to represent data for a specific time period as defined by experts. Association rule mining is applied to analyze and detect abnormal customer behavior. The results of this article indicate that the system is relatively effective in detecting customer churn early on and thus helpful at assisting banks to address issues before they escalate. Furthermore, the tested rules are further scrutinized by experts to establish the relationship between the defined rules and management. This study provides an expert system for banks to assess the quality of their marketing campaigns and reestablish faltering customer relationships.

Prediction of stocks is complicated by the dynamic, complex, and chaotic environment of the stock market. Many studies predict stock price movements using deep learning models. Although the attention mechanism has gained popularity recently in neural machine translation, little focus has been devoted to attention-based deep learning models for stock prediction. This paper proposes an attention-based long short-term memory model to predict stock price movement and make trading strategies.

The massive amounts of data available on social media platforms become the key source of information related to customer sentiment and opinions for analysis by companies. Facebook is one of the most popular platforms which a firm can take advantage of to establish customer relationships through the use of fan pages. Facebook fan pages contain rich sources of information related to customer relationships. This study proposes a novel recommender system based on analyzing the behavior of fans. The proposed method is able to overcome the limitations of traditional recommender system. Three common measures will be used to evaluate the effectiveness of the proposed method. The experimental data have been obtained by crawling from IMDb fan pages..

Online customer reviews of both products and merchants can greatly affect the readers' purchase decision making. Opinion spamming refers to the writing of fake reviews that try to deliberately mislead the reader by giving undeserved positive opinions or unjust or false negative opinions to promote or demote the target product. This paper focuses on analyzing spammer behaviors on a well-known Taiwan web forum, Mobile01. We explore the characteristics of the spam and the spammers operating within such a web forum to obtain some insights for future study.

This paper presents a novel active damping control method for the driving circuit of a piezoelectric transducer (PT) applied in a distance measurement system. Since the vibration of the PT cannot come to a halt instantly after the driving circuit stops driving it, the minimum detectable length in the system with a single PT is restricted. By applying active damping, the transition time can be reduced. Conventionally, a peak detection circuit is needed to realize the active damping control. However, with the proposed control, the phase of the vibration can be known from the driving signal, and the active damping control can be realized without any sensing circuits. The proposed control method is verified with a bi-directional flyback converter. The experimental results show that the amplitude of terminal voltage of the PT can be reduced to one-fourth of the value with the conventional passive damping.

Collaborative Filtering (CF) has been applied to many commercial systems successfully, such as IMDB, Netflix and son on. The basic idea of a CF system is to generate recommendations based on the experiences of past similar users. The users' option can be categorized into objective and subject information. The former was furnished by the common users and the later represents solicit opinions provided by experts (such as film critics). Both information types are valuable and important for the CF system. This study attempts to propose a novel collaborative filtering framework based on fuzzy set theory which integrates the subjective and objective information. The new methodology not only provides a comprehensive result but also solve the problems of traditional CF system, new user and new item. Finally, an experiment is performed, and the result indicates that the proposed methodology produces high-quality recommendations.