Reviewing the Simple Things- How Ease of Evaluation Affects Online Rating Behavior

2020 | European Conference On Information Systems | Citations: 0

Authors: Seutter, Janina; Neumann, Jürgen

Abstract: Online reviews play a considerable role in reducing the information asymmetry b ... Expand

Abstract: Online reviews play a considerable role in reducing the information asymmetry between sellers and potential consumers. Despite the rich body of literature on online reviews and rating behaviors, little is known about the influence that the ease (or difficulty) involved in evaluating a product or service has on ratings. Similarly, certain product or service characteristics are easier to review and evaluate than others. In this paper we investigate the potential rating differences that arise from the ease of evaluation, both on a category and on an attribute level. In two distinct studies we analyze datasets from, respectively, Yelp (category level) and Google Maps (attribute level) and conduct linear regression with fixedeffects. Our results suggest that ratings of easy-to-evaluate product and service categories are more extreme than ratings of difficult-to-evaluate categories, whereas this does not hold for attributes. These contrasting results between category and attribute level reveal that the impact of ease of evaluation cannot be fully explained by expectation-disconfirmation theory and, therefore, we briefly discuss alternative theoretical explanations. Our results also have important practical implications for platforms offering goods and services that differ in ease of evaluation, in order to redress the biases created by these differences in rating behaviors. Collapse

Topics: online review Google+ user-generated content customer satisfaction

Methods: computational algorithm ordinary least square random forest classification sentiment analysis descriptive statistic

Theories: expectation disconfirmation theory

Using similarity measures for medical event sequences to predict mortality in trauma patients

2019 | Decision Support Systems | Citations: 1

Authors: Fredrickson, Joel; Mannino, Michael; Alqahtani, Omar; Banaei-Kashani, Farnoush

Abstract: We extend a similarity measure for medical event sequences (MESs) and evaluate i ... Expand

Abstract: We extend a similarity measure for medical event sequences (MESs) and evaluate its classification performance for retrospective mortality prediction of trauma patient outcomes. Retrospective mortality prediction is a benchmarking task used by trauma care governance bodies to assist with policy decisions. We extend the similarity measure, the Optimal Temporal Common Subsequence for MESs (OTCS-MES), by generalizing the event-matching component with a plug-in weighting element. The extended OTCS-MES uses an event prevalence weight developed in our previous study and an event severity weight developed for this study. In the empirical evaluation of classification performance, we provide a more complete evaluation than previous studies. We compare the predictive performance of the Trauma Mortality Prediction Model (TMPM), an accepted regression approach for mortality prediction in trauma data, to nearest neighbor algorithms using similarity measures for MESs. Using a data set from the National Trauma Data Bank, our results indicate improved predictive performance for an ensemble of nearest neighbor classifiers over TMPM. Our analysis reveals a superior Receiver Operating Characteristics (ROC) curve, larger AUC, and improved operating points on a ROC curve. We also study methods to adjust for uncommon class prediction: weighted voting, neighborhood size, and case base size. Results provide strong evidence that similarity measures for medical event sequences are a powerful and easily adapted method assisting with health care policy advances. Collapse

Topics: missing data

Methods: machine learning classification method archival research experiment regression analysis method

Mechanisms for Automatic Training Data Labeling for Machine Learning

2019 | International Conference on Information Systems | Citations: 4

Authors: Gu, Yang; Leroy, Gondy

Abstract: One of the most pervasive challenges in adopting machine or deep learning is the ... Expand

Abstract: One of the most pervasive challenges in adopting machine or deep learning is the scarcity of training data. This problem is amplified in IS research, where application domains usually require specialized knowledge. This study compares three systems to create a large dataset for training when only a small amount of human-labeled data is available: a high-precision LSTM classifier, a high-recall LSTM classifier, and manually created rule-based system. Based on fewer than 20,000 human-labeled training examples, we used automated labeling to add an additional 100,000 examples to the training data. We found that combining a small human-labeled dataset with a system-labeled dataset improves classification performance. In our evaluation, adding training data labeled by the high-recall LSTM to the human-labeled dataset achieved F1 of 0.578, and adding training data labeled by the rule-based system achieved F1 of 0.598, over 4% improvement compared to a baseline system that only uses human-labeled data. Collapse

Topics: IT productivity health information system Python blog electronic health record

Methods: support vector machine machine learning deep learning computational algorithm voting ensemble

Feature enrichment through multi-gram models

2018 | HICSS | Citations: 0

Authors: Forss, Thomas

Abstract: We introduce a feature enrichment approach, by developing multi-gram cosine simi ... Expand

Abstract: We introduce a feature enrichment approach, by developing multi-gram cosine similarity classification models. Our approach combines cosine similarity features of different N-gram word models, and unsupervised sentiment features, into models with a richer feature set than any of the approaches alone can provide. We test the classification models using different machine learning algorithms on categories of hateful and violent web content, and show that our multi-gram models give across-the-board performance improvements, for all categories tested, compared to combinations of baseline unigram, N-gram, and sentiment classification models. Our multi-gram models perform significantly better on highly imbalanced sets than the comparison methods, while this enrichment approach leaves room for further improvements, by adding instead of exhausting optimization options. Collapse

Topics: web page email spam big data

Methods: machine learning computational algorithm support vector machine experiment k-nearest neighbor

Mortality Prediction using Similarity Measures for Medical Event Sequences

2018 | Americas Conference on Information Systems | Citations: 0

Authors: Fredrickson, Joel; Alqahtani, Omar; Mannino, Michael; Banaei-Kashani, Farnoush

Abstract: We extend a similarity measure for medical event sequences (MESs) and evaluate i ... Expand

Abstract: We extend a similarity measure for medical event sequences (MESs) and evaluate its performance on mortality prediction using a substantial trauma data set. We extend the Optimal Temporal Common Subsequence for MESs (OTCS-MES) measure by generalizing the event-matching component with userdefined weights. In the empirical evaluation of classification performance, we provide a more complete evaluation than previous studies. We compare the predictive performance of the Trauma Mortality Prediction Model (TMPM), an accepted regression approach for mortality prediction in trauma data, to nearest neighbor algorithms using similarity measures for MESs. Using a data set from the National Trauma Data Bank, our results indicate improved predictive performance for an ensemble of nearest neighbor classifiers over TMPM. Our analysis demonstrates a superior Receiver Operating Characteristics (ROC) curve, larger AUC, and improved operating points on a ROC curve. Predictive performance improves for the ensemble for a variety of sensitivity weights and false positive constraints. Collapse

Topics: decision making

Methods: machine learning experiment computational algorithm regression analysis method classification method

Lending Motivation Meets Home and Cultural Bias: A Study on Kiva

2017 | International Conference on Information Systems | Citations: 1

Authors: Yan, Jiaqi; Liu, Yi; Shi, Yani; Wang, Kaixin

Abstract: Online pro-social crowd funding platforms facilitate borrowers from economically ... Expand

Abstract: Online pro-social crowd funding platforms facilitate borrowers from economically weaker locations to access financial supports from lenders globally. However, geographical and cultural distances are still prevailing on these platforms. Loan transactions are more likely to occur between lenders and borrowers in the same country and lenders prefer to fund culturally similar and geographically proximate borrowers. Since, it prevents the borrowers from underrepresented populations to access sufficient loan funding, how to overcome this bias needs to be examined. This study aims to investigate how lenders’ lending motivations affect their lending decisions on the loan requests regarding the regional and cultural differences between lenders and borrowers. We contextualize our study at Kiva.org and lending transactions are retrieved via its API. Lenders’ motivations are mined and categorized based on lenders’ self-stated descriptions. This study potentially contributes to enrich the literature of cross-cultural studies in social lending by studying the effects of lending motivation. Collapse

Topics: decision support crowdsourcing Twitter Python recommender system

Methods: support vector machine classification method word embedding regression analysis method linear discriminant analysis

Distributed Privacy-Preserving Decision Support System for Highly Imbalanced Clinical Data

2013 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Mathew, George; Obradovic, Zoran

Abstract: When a medical practitioner encounters a patient with rare symptoms that transla ... Expand

Abstract: When a medical practitioner encounters a patient with rare symptoms that translates to rare occurrences in the local database, it is quite valuable to draw conclusions collectively from such occurrences in other hospitals. However, for such rare conditions, there will be a huge imbalance in classes among the relevant base population. Due to regulations and privacy concerns, collecting data from other hospitals will be problematic. Consequently, distributed decision support systems that can use just the statistics of data from multiple hospitals are valuable. We present a system that can collectively build a distributed classification model dynamically without the need of patient data from each site in the case of imbalanced data. The system uses a voting ensemble of experts for the decision model. The imbalance condition and number of experts can be determined by the system. Since only statistics of the data and no raw data are required by the system, patient privacy issues are addressed. We demonstrate the outlined principles using the Nationwide Inpatient Sample (NIS) database. Results of experiments conducted on 7,810,762 patients from 1050 hospitals show improvement of 13.68% to 24.46% in balanced prediction accuracy using our model over the baseline model, illustrating the effectiveness of the proposed methodology. Collapse

Topics: database system healthcare data privacy electronic health record decision support

Methods: experiment computational algorithm decision tree classification ensemble learning machine learning