Classification results of the studies analyzed in A State-of-the-Art Review to Examine the Impact of Intelligent Document Processing in Banking Automations

This spreadsheet presents the meticulously classified results from the conducting phase of our systematic literature review titled "From Manual to Automated: A State-of-the-Art Review to Examine the Impact of Intelligent Document Processing in Banking Automation". Each entry within this document represents an individual study analyzed during our research, categorized according to a carefully designed classification framework to ensure a comprehensive and clear understanding of the evolving landscape in banking automation through Intelligent Document Processing (IDP) technologies. Classification Framework Overview RQ1. General Study Characterization Date: indicates the year of publication of the study. Contribution Source: refers to the type of publication in which the study appears, such as a journal article or conference paper. Validation: describes the context in which the study’s findings are validated, distinguishing between research environments and industrial or practical applications. Contribution Type: defines the nature of the study’s main contribution, whether it presents an algorithm, a theoretical analysis, a framework, a method, or a model. Public Data Exposure: reflects whether the study generates original datasets and makes them publicly accessible, distinguishing between contributions that provide new open data and those that rely on existing sources or do not disclose their data. RQ2. Machine Learning Approaches and Trends Learning Paradigm: classifies the study’s learning approach as supervised or unsupervised. AI Subfield: identifies the primary Artificial Intelligence (AI) domain of the study, such as data mining, computer vision, or natural language processing (NLP). Model Category: describes the specific type of Machine Learning (ML) model applied in the study, including rule-based models, regression models, clustering, support vector machines, decision trees, or neural networks. RQ3. Business Automation Strategies Automation Compatibility: assesses whether the study’s proposal aligns with Robotic Process Automation (RPA) or fits within a broader, more general automation context. IDP Life Cycle Stage: defines the phase of the IDP life cycle addressed by the study, such as preprocessing, data extraction, or classification. Business Environment Integration: assesses whether the proposed solution is designed for integration within business environments or remains conceptual. Data Preparation Techniques: describes the preprocessing steps applied to structure and enhance raw inputs, employed in the study, including cleaning, transformation, vectorization, or token and label manipulation. RQ4. Application Areas Application Domain: identifies the sector or industry targeted by the study, such as banking, finance, fraud detection, accounting, or auditing. Case Study: specifies the particular application context or document type addressed by the study, for example, checks, invoices, signatures, or broader document categories. This classification scheme is instrumental in providing a structured, in-depth analysis of the field's current state, trends, and future directions. The framework aids in navigating the vast amount of information in the domain, offering researchers, practitioners, and policymakers a clear vision of the significant aspects of each study to foster informed decisions and further innovation in banking automations through IDP.