Modern Centaurs: How Humans and AI Systems Interact in Sales Forecasting

2023 | European Conference On Information Systems | Citations: 0

Authors: Beese, Jannis; Fahse, Tobias

Abstract: Recent achievements of artificial intelligence (AI) have caused organizations t ... Expand

Abstract: Recent achievements of artificial intelligence (AI) have caused organizations to increasingly bring AI capabilities into their core business processes. Such AI-supported business processes often result in human-AI hybrid systems, which consist of an AI system, which performs most of the execution, and humans, who monitor this execution and occasionally provide additional inputs and overrides. Using sales data from Walmart, we conduct an online study to investigate if human supervision can improve upon state-of-the-art AI forecasts. Furthermore, we analyze the perceptions and behavioral intentions of the human participants over time. We find that human interventions consistently lead to less accurate forecasts and that participants initially underestimate the AI system's accuracy and overestimate their own potential to improve upon AI forecasts. However, perceptions quickly shift over the course of the study, causing the participants to perceive the AI system increasingly favorably, which also leads to behavioral changes and better hybrid system performance. Collapse

Topics: artificial intelligence IT productivity information system use task productivity perceived usefulness

Methods: regression analysis method survey LightGBM conceptual modelling experiment

Theories: technology acceptance model

Can Conversations on Reddit Forecast Future Economic Uncertainty? An Interpretable Machine Learning Approach

2023 | International Conference on Information Systems | Citations: 0

Authors: Jiang, Jinhang; Feng, Mei; Srinivasan, Karthik

Abstract: In recent years, social media has become an indispensable source of information ... Expand

Abstract: In recent years, social media has become an indispensable source of information through which public attitudes, opinions, and concerns can be studied and quantified. This paper proposes an interpretable machine learning framework for predicting the Equity Market-related Economic Uncertainty Index using features generated from a popular discussion forum on Reddit. Our framework consists of a series of custom preprocessing and analytics methods, including BERTopic for latent topic identification and regularized linear models. Using our framework, we evaluate explanatory models with different configurations over a large corpus of Reddit posts belonging to the personal finance category. Our analysis generates valuable insights about discussion topics on Reddit and their efficacy in accurately predicting future economic uncertainty. The study demonstrates the potential of using social media data and interpretable machine learning to inform economic forecasting research. Collapse

Topics: Reddit social media explainable artificial intelligence Python permutation feature importance

Methods: descriptive statistic time series analysis longitudinal research topic model experiment

Improved healthcare disaster decision-making utilizing information extraction from complementary social media data during the COVID-19 pandemic

2023 | Decision Support Systems | Citations: 0

Authors: Kellner, Domenic; Lowin, Maximilian; Hinz, Oliver

Abstract: Managing an extreme event like a healthcare disaster requires accurate informati ... Expand

Abstract: Managing an extreme event like a healthcare disaster requires accurate information about the event's circum­ stances to comprehend the full consequences of acting. However, information quality is rarely optimal since it takes time to determine the information of relevance. The COVID-19 pandemic showed that even official data sources are far from optimal since they suffer from reporting delays that slow decision-making. To support decision-makers with timely information, we utilize data from online social networks to propose an adaptable information extraction solution to create indices helping to forecast COVID-19 case numbers and hospitalization rates. We show that combining heterogeneous data sources like Twitter and Reddit can leverage these sources' inherent complementarity and yield better predictions than those using a single data source alone. We further show that the predictions run ahead of the official COVID-19 incidences by up to 14 days. Additionally, we highlight the importance of model adjustments whenever new information becomes available or the underlying data changes by observing distinct changes in the presence of specific symptoms on Reddit. Collapse

Topics: Twitter Reddit social media pandemic website

Methods: BERT information extraction computational algorithm autoregressive integrated moving average time series analysis

Estimating Life Cycle Sales of Technology Products with Frequent Repeat Purchases: A Fractional Calculus-Based Approach

2023 | Information Systems Research | Citations: 0

Authors: Lotfi, Aslan; Jiang, Zhengrui; Lotfi, Ali; Jain, Dipak C.

Abstract: Accurately predicting the sales trajectory of a product in its life cycle is cri ... Expand

Abstract: Accurately predicting the sales trajectory of a product in its life cycle is critically important for firms’ medium- and long-term planning. Because classic product-diffusion models such as the Bass model consider only initial product purchases, they are ill-fitted for sales prediction for today’s technology products with a shorter life cycle and frequent repeat purchases or subscription renewals. Despite the long tradition of product diffusion research, there exists no viable model option when repeat purchases constitute a large proportion of product sales. The present study introduces a new sales growth model, termed the generalized diffusion model with repeat purchases (GDMR), to fill this void. The GDMR formulates the growth rate of sales using a noninteger-order integral equation rather than the integer-order differential equation typically adopted in existing diffusion models. The GDMR is parsimonious and easy to implement. Empirical results show that the GDMR fits sales data with varying proportions of repeat purchases quite well, making it suitable for predicting sales of a wide variety of products. In addition, the GDMR can be extended to incorporate marketing mix variables, thus enhancing its applicability in business decision making. Furthermore, using both real and simulated data, we show that the GDMR can reliably recover a product’s adoption trend using only sales data, thus cementing its theoretical validity and empirical effectiveness. Finally, we show that the GDMR is superior to generic time series and machine learning models in predicting future product sales. History: Ram Gopal, Senior Editor; Wenjing Duan, Associate Editor. Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2022.1131. Collapse

Topics: marketing management DVD innovation management diffusion of innovation tablet

Methods: longitudinal research time series analysis machine learning simulation autoregressive integrated moving average

Human–Artificial Intelligence Collaboration in Prediction: A Field Experiment in the Retail Industry

2023 | Journal of Management Information Systems | Citations: 0

Authors: Revilla, Elena; Saenz, Maria Jesus; Seifter, Matthias; Ma, Ye

Abstract: This study investigates the role of human intervention in artificial intelligenc ... Expand

Abstract: This study investigates the role of human intervention in artificial intelligence/machine learning (AIML)-driven predictions. By doing so, we distinguish between three different types of human-AIML collaboration: automation, adjustable automation, and augmentation. We theorize that prediction uncertainty and time horizon represent two critical determinants of forecast accuracy. Based on a field experiment involving AIML-driven demand forecasts approximately 1,888 stock-keeping units in the retail industry, we rely on a multivalued treatment effect methodology to measure the effects of human-AIML collaboration on forecast accuracy. Our findings show that human intervention complements AIML-driven forecasts most effectively (augmentation) in predictions with long time horizons and low uncertainty. However human intervention is least likely to contribute to the effectiveness of AIML predictions (automation) in environments with short time horizons and high uncertainty. We discuss implications for extant theory and propose a framework outlining the conditions in which human intervention is most likely to add predictive value to human-AIML collaborations. Collapse

Topics: artificial intelligence human computer interaction human AI collaboration decision making task uncertainty

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

Getting Personal: A Deep Learning Artifact for Text-Based Measurement of Personality

2023 | Information Systems Research | Citations: 0

Authors: Yang, Kai; Lau, Raymond Y. K.; Abbasi, Ahmed

Abstract: Analysts, managers, and policymakers are interested in predictive analytics capa ... Expand

Abstract: Analysts, managers, and policymakers are interested in predictive analytics capable of offering better foresight. It is generally accepted that in forecasting scenarios involving organizational policies or consumer decision making, personal characteristics, including personality, may be an important predictor of downstream outcomes. The inclusion of personality features in forecasting models has been hindered by the fact that traditional measurement mechanisms are often infeasible. Text-based personality detection has garnered attention because of the public availability of digital textual traces. However, the text machine learning space has bifurcated into two branches: feature-based methods relying on manually crafted human intuition, or deep learning language models that leverage big data and compute, the main commonality being that neither branch generates accurate personality assessments, thereby making personality measures infeasible for downstream forecasting applications. In this study, we propose DeepPerson, a design artifact for text-based personality detection that bridges these two branches by leveraging concepts from relevant psycholinguistic theories in conjunction with advanced deep learning strategies. DeepPerson incorporates novel transfer learning and hierarchical attention network methods that use psychological concepts and data augmentation in conjunction with person-level linguistic information. We evaluate the utility of the proposed artifact using an extensive design evaluation on three personality data sets in comparison with state-of-the-art methods proposed in academia and industry. DeepPerson can improve detection of personality dimensions by 10–20 percentage points relative to the best comparison methods. Using case studies in the finance and health domains, we show that more accurate text-based personality detection can translate into significant improvements in downstream applications such as forecasting future firm performance or predicting pandemic infection rates. Our findings have important implications for research at the intersection of design and data science, and practical implications for managers focused on enabling, producing, or consuming predictive analytics.History: Olivia Sheng, Senior Editor; Huimin Zhao, Associate Editor.Funding: This work was supported by Oracle for Research (NLP for the Greater Good) and the U.S. National Science Foundation’s Division of Information and Intelligent Systems [Grants BDS-1636933 and IIS-1816504]. The work was also partly supported by the Research Grants Council of the Hong Kong Special Administrative Region [Project: CityU 11507219] and the City University of Hong Kong SRG [Project: 7005780].Supplemental Material: The online appendices are available at https://doi.org/10.1287/isre.2022.1111. Collapse

Topics: personality Twitter social media pandemic Google+

Methods: machine learning BERT design science deep learning computational algorithm

Theories: big five model

Read the News, Not the Books: Forecasting Firms’ Long-term Financial Performance via Deep Text Mining

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

Authors: Zhai, Shuang (Sophie); Zhang, Zhu (Drew)

Abstract: In this paper, we show textual data from firm-related events in news articles ca ... Expand

Abstract: In this paper, we show textual data from firm-related events in news articles can effectively predict various firm financial ratios, with or without historical financial ratios. We exploit state-of-the-art neural architectures, including pseudo-event embeddings, Long Short-Term Memory Networks, and attention mechanisms. Our news-powered deep learning models are shown to outperform standard econometric models operating on precise accounting historical data. We also observe forecasting quality improvement when integrating textual and numerical data streams. In addition, we provide in-depth case studies for model explainability and transparency. Our forecasting models, model attention maps, and firm embeddings benefit various stakeholders with quality predictions and explainable insights. Our proposed models can be applied both when numerically historical data is or is not available. Collapse

Topics: missing data accounting financial performance sensemaking organizational productivity

Methods: deep learning autoregressive integrated moving average longitudinal research time series analysis case study

Modeling renewable energy production and CO2 emissions in the region of Adrar in Algeria using LSTM neural networks

2022 | HICSS | Citations: 0

Authors: Bouziane, Seif Eddine; Dugdale, Julie; Khadir, Mohamed Tarek

Abstract: This paper addresses the slow-onset crisis of global warming caused by CO2 emiss ... Expand

Abstract: This paper addresses the slow-onset crisis of global warming caused by CO2 emissions. Although electrical load is a major influence in a country’s growth and development, it is also one of largest sources of greenhouse gases (GHG), CO2 in particular. Therefore, switching to cleaner energy sources is a clear objective and forecasting electricity load and its environmental cost is a necessary task for electrical energy planning and management. This paper addresses short-term load forecasting of renewable energy (RE) production in the region of Adrar in Algeria with Adrar’s photovoltaic (PV) farm and Kabertene’s wind farm. The forecast is compared to the overall load demand, and the reduced amount of CO2 resulting from using renewable energy instead of fossil fuels is calculated. The forecasting models use Long short-term memory (LSTM) neural networks, which were trained and validated using real data provided by the national state-owned company SONALGAZ. The results show good performance for the forecasting models with PV and wind models achieving a Mean-absolute-error (MAE) of 0.024 and 0.1 respectively. This RE can help to reduce CO2 emissions by up to 25% per hour. Collapse

Topics: climate change

Methods: autoregressive integrated moving average experiment computational algorithm case study artificial neural network

Explanation Interfaces for Sales Forecasting

2022 | European Conference On Information Systems | Citations: 1

Authors: Fahse, Tobias; Blohm, Ivo; Hruby, Richard; Giffen, Benjamin van

Abstract: Algorithmic forecasts outperform human forecasts in many tasks. State-of-the-ar ... Expand

Abstract: Algorithmic forecasts outperform human forecasts in many tasks. State-of-the-art machine learning (ML) algorithms have even widened that gap. Since sales forecasting plays a key role in business profitability, ML based sales forecasting can have significant advantages. However, individuals are resistant to use algorithmic forecasts. To overcome this algorithm aversion, explainable AI (XAI), where an explanation interface (XI) provides model predictions and explanations to the user, can help. However, current XAI techniques are incomprehensible for laymen. Despite the economic relevance of sales forecasting, there is no significant research effort towards aiding non-expert users make better decisions using ML forecasting systems by designing appropriate XI. We contribute to this research gap by designing a model-agnostic XI for laymen. We propose a design theory for XIs, instantiate our theory and report initial formative evaluation results. A real-world evaluation context is used: A medium-sized Swiss bakery chain provides past sales data and human forecasts. Collapse

Topics: decision quality explainable artificial intelligence artificial intelligence organizational value data literacy

Methods: design theory design principle design requirement design artifact longitudinal research

Theories: cognitive load theory theory of abduction

Effectiveness of Example-Based Explanations to Improve Human Decision Quality in Machine Learning Forecasting Systems

2022 | International Conference on Information Systems | Citations: 0

Authors: Fahse, Tobias; Blohm, Ivo; Giffen, Benjamin van

Abstract: Algorithmic forecasts outperform human forecasts by 10% on average. State-of-th ... Expand

Abstract: Algorithmic forecasts outperform human forecasts by 10% on average. State-of-the-art machine learning (ML) algorithms have further expanded this discrepancy. Because a variety of other activities rely on them, sales forecasting is critical to a company's profitability. However, individuals are hesitant to use ML forecasts. To overcome this algorithm aversion, explainable artificial intelligence (XAI) can be a solution by making ML systems more comprehensible by providing explanations. However, current XAI techniques are incomprehensible for laymen, as they impose too much cognitive load. We contribute to this research gap by investigating the effectiveness in terms of forecast accuracy of two example-based explanation approaches. We conduct an online experiment based on a two-by-two between-subjects design with factual and counterfactual examples as experimental factors. A control group has access to ML predictions, but not to explanations. We report results of this study: While factual explanations significantly improved participants' decision quality, counterfactual explanations did not. Collapse

Topics: decision quality artificial intelligence explainable artificial intelligence mobile application production planning information system

Methods: experimental group longitudinal research experiment time series analysis parametric test

Theories: cognitive load theory