On Predicting ESG Ratings Using Dynamic Company Networks

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Ang, Gary; Guo, Zhiling; Lim, Ee-Peng

Abstract: Environmental, social and governance (ESG) considerations play an increasingly i ... Expand

Abstract: Environmental, social and governance (ESG) considerations play an increasingly important role due to the growing focus on sustainability globally. Entities, such as banks and investors, utilize ESG ratings of companies issued by specialized rating agencies to evaluate ESG risks of companies. The process of assigning ESG ratings by human analysts is however laborious and time intensive. Developing methods to predict ESG ratings could alleviate such challenges, allow ESG ratings to be generated in a more timely manner, cover more companies, and be more accessible. Most works study the effects of ESG ratings on target variables such as stock prices or financial fundamentals of companies, but few works study how different types of company information can be utilized to predict ESG ratings. Previous works also largely focus on using only the financial information of individual companies to predict ESG ratings, leaving out the different types of inter-company relationship networks. Such inter-company relationship networks are typically dynamic, i.e., they evolve across time. In this paper, we focus on utilizing dynamic inter-company relationships for ESG ratings prediction, and examine the relative importance of different financial and dynamic network information in this prediction task. Our analysis shows that utilizing dynamic inter-company network information, based on common director, common investor and news event-based knowledge graph relationships, can significantly improve ESG rating prediction performance. Robustness checks over different time-periods and different number of time-steps in the future further validate these insights. Collapse

Topics: shapley additive explanation knowledge graph database system missing data network effect

Methods: machine learning econometric modeling longitudinal research Granger causality robustness check

Selecting the Optimal Number of Crowd Workers for Forecasting Tasks

2023 | Americas Conference on Information Systems | Citations: 0

Authors: Carvalho, Arthur; Karimi, Majid

Abstract: Crowdsourcing platforms enable requesters to elicit information from thousands ... Expand

Abstract: Crowdsourcing platforms enable requesters to elicit information from thousands of workers worldwide. A question that arises when outsourcing a task to a crowd is how many crowd workers a requester should hire. Focusing on forecasting tasks, we provide a methodological way of answering that question by formally describing how the number of crowd workers relates to the expected forecast error. Specifically, we discuss how an error curve associates a number of hired crowd workers with a fixed and a variable error. We then suggest an optimality concept that enables requesters to select an optimal number of crowd workers based on the tolerable variable error. In an illustrative study on Amazon Mechanical Turk, we demonstrate how highly prescriptive our approach is. In particular, we show how requesters can perform ex-ante and ex-post analyses by determining the optimal number of workers before running a task and by determining the fixed error after running the task and collecting forecasts. Collapse

Topics: crowdsourcing decision support system

Methods: computational algorithm post analysis optimization model experiment mathematical model

The Impact of Social Reputation Features in Innovation Tournaments: Evidence from a Natural Experiment

2023 | Information Systems Research | Citations: 0

Authors: Deodhar, Swanand J.; Gupta, Samrat

Abstract: This study examines how a change in an online reputation system, specifically th ... Expand

Abstract: This study examines how a change in an online reputation system, specifically the addition of a social reputation feature, affects contestant performance in innovation tournaments. Drawing from the literature on peer recognition and social evaluation anxiety, we project competing effects whereby the feature could either enhance or diminish contestant performance. Furthermore, we hypothesize a series of contingent effects involving the soft reserve, a competitive dynamic that unfolds in tournaments, and a determinant of performance in its own right. Specifically, we hypothesize that the direct influence of the social reputation feature on contestant performance would be predicated on the level of two types of soft reserves in an innovation tournament: that created by the focal contestant and that created by competitors. We test these hypotheses leveraging a natural experiment, where an innovation tournament platform (Kaggle.com) introduced a social reputation feature, allowing contestants to follow other contestants unilaterally. Estimates obtained using a panel data set bracketed within a narrow time window (15 days) around the feature launch reveal that the feature significantly improves the performance. We further report that the two types of soft reserves significantly moderate the positive effect of the social reputation feature on contestant performance, whereby the higher the soft reserve, the weaker the effect of the social reputation feature on contestant performance. These findings have several theoretical and practical implications for managing innovation tournaments.History: Manju Ahuja, Senior Editor; Maha Shaikh, Associate Editor.Funding: The authors acknowledge Research & Publications office of IIM Ahmedabad for providing funding for language editing services.Supplemental Material: The online appendices are available at https://doi.org/10.1287/isre.2022.1118. Collapse

Topics: innovation management online reputation system word of mouth accounting blog

Methods: natural experiment robustness check statistical hypothesis test descriptive statistic computational algorithm

Incorporating Multiple Knowledge Sources for Targeted Aspect-based Financial Sentiment Analysis

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Du, Kelvin; Xing, Frank; Cambria, Erik

Abstract: Combining symbolic and subsymbolic methods has become a promising strategy as re ... Expand

Abstract: Combining symbolic and subsymbolic methods has become a promising strategy as research tasks in AI grow increasingly complicated and require higher levels of understanding. Targeted Aspect-based Financial Sentiment Analysis (TABFSA) is an example of such complicated tasks, as it involves processes like information extraction, information specification, and domain adaptation. However, little is known about the design principles of such hybrid models leveraging external lexical knowledge. To fill this gap, we define anterior, parallel, and posterior knowledge integration and propose incorporating multiple lexical knowledge sources strategically into the fine-tuning process of pre-trained transformer models for TABFSA. Experiments on the Financial Opinion mining and Question Answering challenge (FiQA) Task 1 and SemEval 2017 Task 5 datasets show that the knowledge-enabled models systematically improve upon their plain deep learning counterparts, and some outperform state-of-the-art results reported in terms of aspect sentiment analysis error. We discover that parallel knowledge integration is the most effective and domain-specific lexical knowledge is more important according to our ablation analysis. Collapse

Topics: microblog domain-specific language Twitter social media knowledge graph

Methods: sentiment analysis RoBERTa transformer model BERT XLNet

A Comparison of Methods for Treatment Assignment with an Application to Playlist Generation

2023 | Information Systems Research | Citations: 0

Authors: Fernández-Loría, Carlos; Provost, Foster; Anderton, Jesse; Carterette, Benjamin; Chandar, Praveen

Abstract: This study presents a systematic comparison of methods for individual treatment ... Expand

Abstract: This study presents a systematic comparison of methods for individual treatment assignment, a general problem that arises in many applications and that has received significant attention from economists, computer scientists, and social scientists. We group the various methods proposed in the literature into three general classes of algorithms (or metalearners): learning models to predict outcomes (the O-learner), learning models to predict causal effects (the E-learner), and learning models to predict optimal treatment assignments (the A-learner). We compare the metalearners in terms of (1) their level of generality and (2) the objective function they use to learn models from data; we then discuss the implications that these characteristics have for modeling and decision making. Notably, we demonstrate analytically and empirically that optimizing for the prediction of outcomes or causal effects is not the same as optimizing for treatment assignments, suggesting that, in general, the A-learner should lead to better treatment assignments than the other metalearners. We demonstrate the practical implications of our findings in the context of choosing, for each user, the best algorithm for playlist generation in order to optimize engagement. This is the first comparison of the three different metalearners on a real-world application at scale (based on more than half a billion individual treatment assignments). In addition to supporting our analytical findings, the results show how large A/B tests can provide substantial value for learning treatment-assignment policies, rather than simply for choosing the variant that performs best on average.History: Olivia Liu Sheng, Senior Editor; Gautam Pant, Associate Editor.Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2022.1149. Collapse

Topics: decision making digital service music streaming service Spotify marketing management

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

Live streaming recommendations based on dynamic representation learning

2023 | Decision Support Systems | Citations: 0

Authors: Gao, Ge; Liu, Hongyan; Zhao, Kang

Abstract: As an emerging form of social media, live streaming services (e.g., Twitch and C ... Expand

Abstract: As an emerging form of social media, live streaming services (e.g., Twitch and Clubhouse) allow users to interact with hosts and peers in real time while enjoying shows or participating in discussions. These platforms are also dynamic, with shows or discussions changing quickly inside a room and users frequently switching between rooms. To improve user engagement and experience on such platforms, we design a new recommendation model named Dynamic Representations for Live Streaming Rooms (DRIVER) to provide room recommendations. Guided by the Integrated Framework for Consumer Path Modeling and the social affordance theory, DRIVER infers dynamic representations of live streaming rooms by leveraging users’ behavior paths in entering, staying in, and leaving rooms. One contribution of our model is a new and efficient dynamic learning framework to model instantaneous and ever-changing inter-room relationships by considering individual users’ behavior paths after leaving a room. Also supported by social affordance theory, another methodological novelty of our model is to capture dynamic characteristics of a room by incorporating features of the current audience inside the room. Experiments on real-world datasets from two different types of live streaming platforms demonstrate that DRIVER outperforms state-of-the-art representation learning methods and sequential recommender systems. The proposed method also has implications for recommender system design in other contexts, in which items are characterized by users’ dynamic behavior paths and ongoing social interactions. Collapse

Topics: live streaming platform recommender system user experience knowledge sharing systems design

Methods: experiment sensitivity analysis ablation analysis descriptive statistic deep learning

Theories: theory of affordance

Green Data Analytics of Supercomputing from Massive Sensor Networks: Does Workload Distribution Matter?

2023 | Information Systems Research | Citations: 0

Authors: Guo, Zhiling; Li, Jin; Ramesh, Ram

Abstract: Energy costs represent a significant share of the total cost of ownership in hig ... Expand

Abstract: Energy costs represent a significant share of the total cost of ownership in high-performance computing (HPC) systems. Using a unique data set collected by massive sensor networks in a petascale national supercomputing center, we first present an explanatory model to identify key factors that affect energy consumption in supercomputing. Our analytic results show that, not only does computing node utilization significantly affect energy consumption, workload distribution among the nodes also has significant effects and could effectively be leveraged to improve energy efficiency. Next, we establish the high model performance using in-sample and out-of-sample analyses. We then develop prescriptive models for energy-optimal runtime workload management and extend the models to consider energy consumption and job performance tradeoffs. Specifically, we present four dynamic resource management methodologies (packing, load balancing, threshold-based switching, and energy optimization), model their application at two levels (purely within-rack and jointly cross-rack resource allocation), and explore runtime resource redistribution policies for jobs under the emergent principle of computational steering and comparatively evaluate strategies that use computational steering with those that do not. Our experimental studies show that packing is preferred when the total workload of a rack is higher than a threshold and load balancing is preferred when it is lower. These results lead to a threshold strategy that yields near-optimal energy efficiency under all workload conditions. We further calibrate the energy-optimal resource allocations over the full range of workloads and present a bicriteria evaluation to consider energy consumption and job performance tradeoffs. We demonstrate significant energy savings of our proposed strategies under various workload conditions. We conclude with implementation guidelines and policy insights into energy-efficient computing resource management in large supercomputing data centers. History: Olivia Liu Sheng, Senior Editor; Xiaobai Li, Associate Editor. Funding: The authors thank the National Supercomputing Center of Singapore for enabling and supporting this research. J. Li received support from the National Natural Science Foundation of China [Grants 71901169, 72032006, and T2293774]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/isre.2023.1208. Collapse

Topics: data center high-performance computing task productivity resource allocation analytical information system

Methods: optimization model experiment simulation descriptive statistic longitudinal research

Cyber threat detection: Unsupervised hunting of anomalous commands (UHAC)

2023 | Decision Support Systems | Citations: 0

Authors: Kayhan, Varol O.; Agrawal, Manish; Shivendu, Shivendu

Abstract: The cyber security industry is rapidly adopting threat hunting as a proactive to ... Expand

Abstract: The cyber security industry is rapidly adopting threat hunting as a proactive tool for early and faster detection of suspected malicious actors. In this paper, we propose a machine learning-based method, Unsupervised Hunting of Anomalous Commands (UHAC), to detect text-based anomalous commands in security information and event management (SIEM) logs that are good candidates for threat hunting. A unique feature of the proposed method is that it first creates a feature set based on the augmentation of document-term and document-character matrices. Then, an autoencoder-based detector is trained on this feature set using a custom loss function. UHAC consis­ tently outperforms other feature sets and algorithms such as one-class support vector machine, density-based spatial clustering of applications with noise, and word-embedding based models such as word2vec. The UHAC detector identifies 84–89% of anomalies in the top 10% of the data. Findings have implications for cybersecurity analysts who perform threat hunting in SIEM logs for process auditing on endpoint devices. Collapse

Topics: IT security artificial intelligence Python decision support system email phishing

Methods: autoencoder data transformation DBSCAN anomaly detection machine learning

Utilizing the omnipresent: Incorporating digital documents into predictive process monitoring using deep neural networks

2023 | Decision Support Systems | Citations: 0

Authors: Levich, Sergej; Lutz, Bernhard; Neumann, Dirk

Abstract: Predictive process monitoring (PPM) allows companies to improve the efficiency o ... Expand

Abstract: Predictive process monitoring (PPM) allows companies to improve the efficiency of their business processes by predicting aspects such as the process outcome, the next event, or the time until the next event. So far, existing studies have mainly focused on developing novel predictive models while using features solely from event logs. In this study, we aim to go beyond log data and increase the focus of PPM research towards external context information. To this end, we consider digital documents as they are omnipresent in many business processes and their inclusion can often be justified by a business rationale. However, incorporating digital documents into PPM models poses considerable challenges as they present unstructured data that can contain visual and textual cues of future process behavior, while manual feature extraction is generally not feasible. Therefore, we propose an approach that processes digital documents based on automated visual and textual feature extraction methods. Furthermore, we design a tailored integration module which transforms the extracted features from multiple document pages into a fixed-size representation that subsequently serves as input for the predictive models. Our evaluation, based on a real-world dataset of insurance claims from a mid-sized German insurance company, featuring 5131 process instances with 32,058 events and 39,242 document pages, shows that incorporating digital documents improves the performance by significant margins in predicting the damage type, the next event, and the time until the next event. Finally, we analyze how digital documents contribute to the model’s predictions in terms of Shapley additive explanations. Collapse

Topics: shapley additive explanation business process management artificial intelligence explainable artificial intelligence

Methods: BERT machine learning descriptive statistic artificial neural network business process modeling

Improved learning performance for small datasets in high dimensions by new dual-net model for non-linear interpolation virtual sample generation

2023 | Decision Support Systems | Citations: 0

Authors: Lin, Liang-Sian; Lin, Yao-San; Li, Der-Chiang; Liu, Yun-Hsuan

Abstract: The number of reliable samples obtained in early decision-making activity is usu ... Expand

Abstract: The number of reliable samples obtained in early decision-making activity is usually relatively small. Due to variable distribution and incomplete structure of tiny datasets, it is challenging to create reliable and robust predictive modeling using classic statistical and machine learning models in small sample settings. The virtual sample generation (VSG) technique improves model learning accuracies for minimal datasets across diverse applications. Virtual samples on independent variables were generated using established VSG methods predi­ cated on the assumption of a probability distribution or a membership function to fill data gaps. However, in the actual world, non-linear function interactions between variables are common. To address this issue, this paper developed a novel VSG method called Dual-VSG, which generates non-linear interpolation virtual samples using a self-supervised learning (SSL) framework to improve learning performance on small datasets. We generated non-linear interpolation virtual samples without labels by estimating non-linear functions and transforming them into a high-dimensional space using the proposed dual-net model. The weights of the dual-net model are transferred to a downstream task to generate virtual sample labels. To demonstrate the effectiveness of the suggested strategy, this research employed five datasets. On the Backpropagation Neural Networks (BPNN) predictive model, we compared the suggested method’s prediction performance to two state-of-the-art VSG approaches. To assess prediction performance on a regression dataset, the Mean Absolute Percentage Error (MAPE) and the Root Mean Square Error (RMSE) are used. Furthermore, the classification accuracy (ACC) and the Fl measure are used to assess classification capability on classification datasets. In addition, the paired t-test was utilized to see if the suggested Dual-VSG approach differed significantly from the other VSG methods in terms of RMSE, MAPE, accuracy (ACC), or F1 score. For short datasets, the suggested Dual-VSG method out­ performs those VSG methods, according to our experimental results. Collapse

Topics: Python programming language

Methods: data transformation t-distributed stochastic neighbor embedding experiment machine learning parametric test