Insegnamento a.a. 2026-2027

Accounting and Institutional Foundations

• The role of accounting information as an economic and social signal.
• Financial statement analysis and its empirical applications.
• Information asymmetries, governance, and disclosure in institutional contexts.

From Econometrics to Machine Learning

• Review of standard empirical methods in accounting research: panel data and regression frameworks.
• Supervised and unsupervised machine learning: principles, model selection, and evaluation.
• From prediction to inference: interpreting ML outputs in an institutional context.

Dynamic and Event-Based Models

• Survival analysis and time-to-event modeling in accounting and finance.
• Multi-state models for studying transitions and corporate events.

Functional and Robust Statistical Methods

• Functional data analysis for longitudinal and curve-valued observations.
• Robust statistics and depth-based measures for handling heterogeneity and outliers.

Natural Language Processing and Text Analysis

• Dictionary-based inference and sentiment analysis of corporate disclosures.
• Topic modeling for large-scale narrative analysis.
• Word embeddings and language models for semantic representation.

Generative AI and Frontier Methods

• Large language models: architecture, capabilities, and limitations.
• Applications of generative AI in accounting, auditing, and financial analysis.
• Responsible use of AI in data-driven decision-making.

Accounting and Institutional Foundations

• Describe the role of accounting information as a structured representation of firm behavior, governance, and communication with markets.
• Identify the principal sources of information asymmetry in financial reporting and explain their implications for investors, auditors, and regulators.
• Distinguish between the informational content of quantitative financial data and unstructured qualitative disclosures.

From Econometrics to Machine Learning

• Explain the conceptual transition from traditional econometric frameworks to machine learning approaches in empirical accounting research.
• Distinguish between supervised and unsupervised learning paradigms and identify appropriate use cases for each in accounting and finance.
• Describe the main criteria for model selection and explain how predictive performance is evaluated and validated.

Dynamic and Event-Based Models

• Explain the principles of survival analysis and describe how time-to-event data differ from standard cross-sectional or panel observations.
• Identify the main applications of multi-state models in the study of corporate transitions and organizational events.

Functional and Robust Statistical Methods

• Describe the concept of functional data and explain when functional data analysis is preferable to standard multivariate methods.
• Explain the rationale for robust statistical methods and identify settings in accounting and finance where standard estimators are likely to underperform.

Natural Language Processing and Text Analysis

• Distinguish between dictionary-based, topic model, and embedding-based approaches to text analysis and explain the assumptions underlying each.
• Describe how NLP techniques can be applied to corporate disclosures to extract signals about firm behavior, strategy, and risk.
• Explain the concept of word embeddings and summarize how language models represent semantic meaning.

Generative AI and Frontier Methods

• Describe the architecture and operating principles of large language models and identify their principal capabilities and limitations.
• Recognize the main ethical and methodological risks associated with the use of generative AI in data-driven research and professional practice.
• Summarize the current and emerging applications of generative AI in accounting, auditing, and financial analysis.

Accounting and Institutional Foundations

• Interpret financial statements and qualitative disclosures to assess firm behavior, governance quality, and information environment.
• Evaluate the empirical relevance of accounting signals in a given institutional or regulatory context.

From Econometrics to Machine Learning

• Apply machine learning pipelines to real-world accounting and financial datasets, from data preparation and feature engineering through model training and evaluation.
• Select the appropriate statistical or machine learning method for a given research or policy question, justifying the choice against plausible alternatives.
• Construct and interpret model performance diagnostics, distinguishing between in-sample fit and genuine out-of-sample predictive ability.

Dynamic and Event-Based Models

• Apply survival analysis techniques to study the timing and determinants of corporate events such as default, auditor switching, or executive turnover.
• Interpret hazard rates and survival curves in an economically meaningful way, connecting statistical outputs to institutional explanations.

Functional and Robust Statistical Methods

• Apply functional data analysis tools to longitudinal accounting data, extracting meaningful patterns from curve-valued observations.
• Implement robust estimation methods when faced with heterogeneous distributions or outlier-contaminated datasets typical of financial data.

Natural Language Processing and Text Analysis

• Process and analyze large corpora of corporate disclosures using dictionary-based, topic model, and embedding-based NLP pipelines.
• Construct text-based measures of firm narrative, sentiment, or strategy and integrate them into empirical accounting research designs.
• Critically assess the validity and reliability of NLP-derived measures, recognizing their assumptions and potential sources of bias.

Generative AI and Frontier Methods

• Employ generative AI tools to support data analysis, code development, and interpretation of empirical results in a responsible and methodologically aware manner.
• Evaluate the outputs of large language models critically, identifying hallucinations, inconsistencies, and domain-specific limitations.

Transversal Skills

• Collaborate effectively in a team to design and execute an end-to-end data analytics project, managing task allocation, version control, and collective presentation of results.
• Communicate complex analytical findings clearly and concisely to audiences with varying levels of technical background, both in written reports and live presentations.

• Lectures
• Collaborative Works / Assignments

• Collaborative Works / Assignments: The primary assessment of the course takes the form of a group data analytics project in which small teams identify a research or policy-relevant question in accounting or finance, assemble and process appropriate data, implement one or more of the computational methods covered in the course, and deliver both a technical written report and a live in-class presentation. The project is designed to replicate the structure of real-world analytical work, requiring students to integrate domain knowledge, methodological rigor, and effective communication within a collaborative setting.

• One group assignment consisting of a technical report and an in-class presentation, representing 80% of final grade. Students apply the methods covered in the course to a realistic dataset or case of their choice, verifying their ability to implement, evaluate, and communicate a data-driven solution — directly assessing the applying-knowledge and transversal (teamwork, communication) outcomes. The in-class discussion also allows the instructor to verify individual understanding through follow-up questions.
• One final multiple-choice written exam representing 20% of final grade, verifying students' acquisition of the core theoretical and conceptual knowledge covered in the course.

• One individual assignment consisting of a technical report and a pre-recorded video presentation, representing 60% of final grade, submitted to the instructor for review. This verifies the same applying-knowledge outcomes as the group assignment, adapted to an individual format since non-attending students do not take part in in-class discussion.
• One final multiple-choice written exam representing 40% of final grade, verifying the same theoretical knowledge outcomes as for attending students; the higher weight compensates for the absence of an in-class discussion component.

Main source

• Slides, handouts, code, and data made available by the instructor on the course platform prior to each session.

---------------------------------------------------------------------------------------------------------------------------------------------------------------------

WHAT FOLLOWS IS A LIST OF ADDITIONAL AND OPTIONAL TEXTBOOKS. SOME OF THEM ARE FREE RESOURCES. 

---------------------------------------------------------------------------------------------------------------------------------------------------------------------

Natural Language Processing and Text Mining

• J. SILGE, D. ROBINSON, Text Mining with R: A Tidy Approach, O'Reilly. Freely available at https://www.tidytextmining.com.
• E. HVITFELDT, J. SILGE, Supervised Machine Learning for Text Analysis in R, O'Reilly. Freely available at https://smltar.com.

Machine Learning

• G. JAMES, D. WITTEN, T. HASTIE, R. TIBSHIRANI, An Introduction to Statistical Learning with Applications in R, Springer. Freely available at https://www.statlearning.com.

Natural Language Processing and Large Language Models

• J. JURAFSKY, J. MARTIN, Speech and Language Processing, 3rd ed., freely available at https://web.stanford.edu/~jurafsky/slp3.

Robust Statistics

• R. MARONNA, R. DOUGLAS MARTIN, V. YOHAI, M. SALIBIÁN-BARRERA, Robust Statistics: Theory and Methods (with R), 2nd ed., Wiley. Available here: https://ndl.ethernet.edu.et/bitstream/123456789/33513/1/Ricardo%20A.%20Maronna_2006.pdf 

Deep Learning and Generative AI

• F. CHOLLET, Deep Learning with R, Manning Publications.

Survival Analysis

• E. VITTINGHOFF, D. GLIDDEN, S. SHIBOSKI, C. MCCULLOCH, Regression Methods in Biostatistics, Springer. — Accessible treatment of survival and multi-state models with applied examples.

Advanced readings

• T. HASTIE, R. TIBSHIRANI, J. FRIEDMAN, The Elements of Statistical Learning, Springer. Freely available at https://web.stanford.edu/~hastie/ElemStatLearn.