An Efficient Deep Learning Paradigm for Deceit Identification Test on EEG Signals

2021 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Edla, Damodar Reddy; Dodia, Shubham; Bablani, Annushree; Kuppili, Venkatanareshbabu

Abstract: Brain-Computer Interface is the collaboration of the human brain and a device th ... Expand

Abstract: Brain-Computer Interface is the collaboration of the human brain and a device that controls the actions of a human using brain signals. Applications of brain-computer interface vary from the field of entertainment to medical. In this article, a novel Deceit Identification Test is proposed based on the Electroencephalogram signals to identify and analyze the human behavior. Deceit identification test is based on P300 signals, which have a positive peak from 300 ms to 1,000 ms of the stimulus onset. The aim of the experiment is to identify and classify P300 signals with good classification accuracy. For preprocessing, a band-pass filter is used to eliminate the artifacts. The feature extraction is carried out using “symlet” Wavelet Packet Transform (WPT). Deep Neural Network (DNN) with two autoencoders having 10 hidden layers each is applied as the classifier. A novel experiment is conducted for the collection of EEG data from the subjects. EEG signals of 30 subjects (15 guilty and 15 innocent) are recorded and analyzed during the experiment. BrainVision recorder and analyzer are used for recording and analyzing EEG signals. The model is trained for 90% of the dataset and tested for 10% of the dataset and accuracy of 95% is obtained. Collapse

Topics: deception detection terrorist incident personality computer hardware

Methods: machine learning electroencephalography autoencoder data transformation experiment

Uplift modeling with value-driven evaluation metrics

2021 | Decision Support Systems | Citations: 0

Authors: Gubela, Robin M.; Lessmann, Stefan

Abstract: Measuring the success of targeted marketing actions is challenging. Research on ... Expand

Abstract: Measuring the success of targeted marketing actions is challenging. Research on prescriptive analytics recom­ mends uplift models to guide targeting decisions. Uplift models predict how much a marketing action will change customers’ behavior, known as the individual treatment effect (ITE). Marketers can then solicit customers in decreasing order of their estimated ITE. We argue that the ITE-based targeting policy is not fully consistent with a business value maximization objective. We propose business-centric evaluation metrics that integrate estimates of the ITE and the expected business value and validate their benefits relative to the ITE-based targeting baseline using real-world marketing data. The new metrics yield remarkably higher profit across different uplift models, targeting depths, profit functions, and data sets. They further contribute to the growing field of interpretable data science by uncovering interdependencies between covariates, ITE, and profit and by clarifying whether cus­ tomers are worth targeting because of high responsiveness or high value. Collapse

Topics: marketing management organizational value decision making electronic mail open source

Methods: random forest classification regression analysis method lasso regression logistic regression linear discriminant analysis

Bankruptcy prediction for SMEs using transactional data and two-stage multiobjective feature selection

2021 | Decision Support Systems | Citations: 0

Authors: Kou, Gang; Xu, Yong; Peng, Yi; Shen, Feng; Chen, Yang; Chang, Kun; Kou, Shaomin

Abstract: Many bankruptcy prediction models for small and medium-sized enterprises (SMEs) ... Expand

Abstract: Many bankruptcy prediction models for small and medium-sized enterprises (SMEs) are built using accountingbased financial ratios. This study proposes a bankruptcy prediction model for SMEs that uses transactional data and payment network–based variables under a scenario where no financial (accounting) data are required. Offline and online test results both confirmed the predictive capability and economic benefit of transactional data–based variables. However, incorporating those features in predictive models produces high dimensional problems, which deteriorates model interpretability and increases feature acquisition costs. Thus, we propose a two-stage multiobjective feature-selection method that optimizes the number of features as well as model classification performance. The results showed that the proposed model achieved similar classification perfor­ mance while greatly reducing the cardinality of the feature subset. Finally, the feature importance evaluation for features in the optimal subset confirmed the importance of transactional data and payment network-based variables for bankruptcy prediction. Collapse

Topics: permutation feature importance accounting information privacy law legal context decision making

Methods: machine learning computational algorithm experiment support vector machine decision tree classification

Deep Learning for Multi-instance Biometric Privacy

2021 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Sudhakar, Tanuja; Gavrilova, Marina

Abstract: The fundamental goal of a revocable biometric system is to defend a user’s biome ... Expand

Abstract: The fundamental goal of a revocable biometric system is to defend a user’s biometrics from being compromised. This research explores the application of deep learning or Convolutional Neural Networks to multi-instance biometrics. Modality features are transformed into revocable templates through the application of random projection. During the user authentication phase, we employ Support Vector Machines, chosen over three other alternative classifiers after carrying out a comparative study. Comparison of the proposed method over other standard deep learning models and performance evaluation before and after revocability have also been discussed. Results demonstrate ability to improve identification accuracy and provide sound template security. The system was validated on three multi-instance iris and fingervein databases. Collapse

Topics: biometric database system privacy electronic authentication application programming interface

Methods: support vector machine deep learning artificial neural network data transformation k-nearest neighbor

Predicting Fall Risks Vulnerability with Inpatient Data in Acute Care Hospitalization

2020 | Americas Conference on Information Systems | Citations: 0

Authors: Ossai, Chinedu; O’Connor, Louise; Wickramasinghe, Nilmini

Abstract: The increasing trend of patients' falls-related impairments in acute care hospi ... Expand

Abstract: The increasing trend of patients' falls-related impairments in acute care hospitalization and the consequent implications for quality care and cost of care have made it prudent to develop techniques for rapid estimation of fall risks on admission. We develop a framework that relies on Extra Tree Classifier (ETC) with class balanced features to model fall risks on admission using risk scores and removing redundant clinical and psychosocial characteristics via multicollinearity and significance testing. The model predicts fall risks on admission to an accuracy of 96.76% and has a higher accuracy of 3.63%-39.32% when compared to Logistics Regression (LR), Linear Discriminant Analysis (LDA), decision tree classifier(DTC), Quadratic Discriminant Analysis(QDA), K Nearest Neighbour(KNN), Support Vector Machine (SVM), Ridge model (RCV) and Artificial Neural Network (ANN). Due to the effectiveness of this model, it is expected that the multifactorial considerations will culminate in cost-effective management and better patient experience. Collapse

Topics: logistics management data accuracy missing data

Methods: computational algorithm machine learning support vector machine decision tree classification artificial neural network

Detecting Political Bots on Twitter during the 2019 Finnish Parliamentary Election

2020 | HICSS | Citations: 0

Authors: Rossi, Sippo; Rossi, Matti; Upreti, Bikesh Raj; Liu, Yong

Abstract: In recent years, the impact of bots used for manipulating public opinion has bec ... Expand

Abstract: In recent years, the impact of bots used for manipulating public opinion has become an increasingly prevalent topic in politics. Numerous sources have reported about the presence of political bots in social media sites such as Twitter. Compared to other countries, the influence of bots in Finnish politics has received little attention from media and researchers. This study aims to investigate the influence of bots on Finnish political Twitter, based on a dataset consisting of the accounts following major Finnish politicians before the Finnish parliamentary election of 2019. To identify the bots, we extend the existing models with the use of user-level metadata and state-of-art classification models. The results support our model as a suitable instrument for detecting Twitter bots. We found that, albeit there is a huge amount of bot accounts following major Finnish politicians, it is unlikely resulting from foreign entities’ attempts to influence the Finnish parliamentary election. Collapse

Topics: Twitter

Methods: computational algorithm random forest classification support vector machine decision tree classification supervised learning

An Exploration of the Drivers of Non- Adoption Behavior: A Discriminant Analysis Approach

2020 | ACM SIGMIS Database | Citations: 0

Authors: Wolverton, Colleen Carraher; Cenfetelli, Ronald

Abstract: While there has been a substantial amount of attention within the information sy ... Expand

Abstract: While there has been a substantial amount of attention within the information systems research community towards understanding the phenomenon of adoption, much less is known about non-adoption. This study examines the factors surrounding the decision to not adopt a technology and whether certain factors exert differing effects on individuals in particular ways such that concurrent factors could be identified to develop a classification of the specific types of non-adoption behavior. Utilizing inhibitor theory and the symbolic adoption model as a foundational framework for the different types of non-adoption, we posit that different types of non-adoption exist which is demonstrated by determining the perceptions towards technology that coalesce around different types of non-adoption. We conducted a two-phase investigation into nonadoption with two goals in mind: 1) identify and explore specific factors of the IT that are associated with the rejection decision and are distinct from the adoption decision; and 2) determine the extent to which these factors (along with traditional enablers) differentiate between different types of non-adoption. The results from a discriminant function analysis (DFA) indicate the coalescence of specific perceptual variables according to the types of non-adoption behavior, specifically, the discriminatory power of differing perceptions of IT between trial rejecters, symbolic rejecters, trial accepters, symbolic adopters, and adopters. The implications for research and implications for practice are discussed. Collapse

Topics: innovation management information overload ease of use Twitter Apple iPhone

Methods: qualitative interview survey data transformation business process modeling linear discriminant analysis

AUTOMATIC ADAPTATION OF EXPOSURE INTENSITY IN VR ACROPHOBIA THERAPY, BASED ON DEEP NEURAL NETWORKS

2019 | European Conference On Information Systems | Citations: 0

Authors: Balan, Oana; Moise, Gabriela; Moldoveanu, Alin; Moldoveanu, Florica; Leordeanu, Marius

Abstract: This paper proposes a real-time Virtual Reality game for treating acrophobia tha ... Expand

Abstract: This paper proposes a real-time Virtual Reality game for treating acrophobia that automatically tailors in-game exposure to heights to the players’ individual characteristics – affective state and physiological features. The elements of novelty are the automatic estimation of fear and the prediction of the next game level based on the electroencephalogram (EEG) and biophysical data – Galvanic Skin Response (GSR) and Heart Rate (HR). Two neural networks have been trained with the data recorded in an experiment where 4 subjects have been in-vivo and virtually exposed to various heights. In order to test the validity of the approach, the same users played the acrophobia game, using two modalities of expressing fear level. After completing a game level, the EEG and biophysical data were averaged and one neural network estimated the current fear score, while the other predicted the next game level. A measure of similarity between the self-estimated fear level during a game epoch and the fear level predicted by the first neural network showed an accuracy rate of 73% and 42% respectively for the two modalities of expressing fear level. 3 out of 4 users succeeded to obtain a fear level of 0 (complete relaxation) in the final game epoch. Collapse

Topics: virtual reality Python database system virtual world gamification

Methods: electroencephalography experiment deep learning artificial neural network Keras

An Exploration of the Drivers of Non-Adoption Behavior: A Discriminant Analysis Approach

2019 | ACM SIGMIS Database | Citations: 0

Authors: Wolverton, Colleen Carraher; Cenfetelli, Ronald

Abstract: While there has been a substantial amount of attention within the information sy ... Expand

Abstract: While there has been a substantial amount of attention within the information systems research community towards understanding the phenomenon of adoption, much less is known about non-adoption. This study examines the factors surrounding the decision to not adopt a technology and whether certain factors exert differing effects on individuals in particular ways such that concurrent factors could be identified to develop a classification of the specific types of non-adoption behavior. Utilizing inhibitor theory and the symbolic adoption model as a foundational framework for the different types of non-adoption, we posit that different types of non-adoption exist which is demonstrated by determining the perceptions towards technology that coalesce around different types of non-adoption. We conducted a two-phase investigation into non-adoption with two goals in mind: 1) identify and explore specific factors of the IT that are associated with the rejection decision and are distinct from the adoption decision, and 2) determine the extent to which these factors (along with traditional enablers) differentiate between different types of non-adoption. The results from a discriminant function analysis (DFA) indicate the coalescence of specific perceptual variables according to the types of non-adoption behavior, specifically, the discriminatory power of differing perceptions of IT between trial rejecters, symbolic rejecters, trial accepters, symbolic adopters, and adopters. The implications for research and implications for practice are discussed. Collapse

Topics: innovation management information overload ease of use Apple iPhone Twitter

Methods: qualitative interview survey data transformation business process modeling discriminant analysis

Impact of Heterogeneous Prior Contribution on Reciprocity in Online Sellers’ Community

2019 | International Conference on Information Systems | Citations: 0

Authors: Yi, Chen

Abstract: This study explores how different types of resources a seller contributes in the ... Expand

Abstract: This study explores how different types of resources a seller contributes in the online community will trigger others’ reciprocity, reflected by the responses the seller’s threads receive. Drawing on social exchange theory and using machine learning techniques, we identify two important types of resources transferred in an online sellers’ community: informational resource and instrumental resource. Our findings reveal that a seller’s provision of informational resource is positively associated with the responses the seller’s threads receive, while a seller’s provision of instrumental resource is negatively associated with the responses the seller’s threads receive. Moreover, the effect is moderated by the types of resources sought by a certain thread. Specifically, both effects are strengthened for threads that seek informational resource and undermined for threads that seek instrumental resource. The study contributes to the understanding of online reciprocity by uncovering the differing impact of different contribution and its boundary condition. Collapse

Topics: online community information exchange electronic commerce

Methods: computational algorithm qualitative coding classification method forums posting robustness check

Theories: social exchange theory