Forecast Performance
Estimate future outcomes and compare scenarios using historical financial information.
Financial analytics
Using data and machine learning to understand risk, performance, and future opportunity.
Financial decisions depend on imperfect information. My work explores how numerical metrics, text, forecasting, and transparent modeling can help decision-makers evaluate evidence more systematically.
Extract Text Signals
Convert SEC filing language into measurable indicators of sentiment, tone, and readability.
Evaluate Risk
Test model stability, sensitivity, uncertainty, and out-of-sample performance.
Improve Transparency
Use interpretable models and coefficient analysis to explain what drives predictions.
Support Investment Research
Compare financial-only, text-only, and combined information pipelines.
Strengthen Decisions
Translate model evidence into practical recommendations without overstating certainty.
Featured case study
Stock Return Prediction Using SEC Filing Text
Testing whether language from company filings provides predictive information beyond traditional financial metrics.
Financial problemDecision-makers need to separate useful signals from market noise.
Analytical responseBuild financial-only, text-only, and combined models.
Validation approachUse chronological splits and a buffer to reduce leakage.
Practical valueProvide transparent evidence about whether text improves prediction.
Model Comparison
Financial-only, text-only, and combined pipelines are compared to determine whether language contributes incremental predictive value.
Feature Interpretation
Coefficient analysis helps explain which financial and textual variables are most influential within the model.
Interpretable Baseline
OLS provides a transparent baseline that supports coefficient interpretation, statistical review, and honest model comparison.
Decision Process
The project moves from research design and feature engineering to validation, interpretation, and financial decision support.
Financial applications beyond this project
Machine learning can support many areas of finance.
Credit Risk
Estimate probability of default and support more consistent lending decisions.
Fraud Detection
Identify unusual transactions and patterns requiring further investigation.
Portfolio Analytics
Analyze risk, diversification, performance, and scenario sensitivity.
Revenue Forecasting
Estimate future revenue and support budgeting and capital planning.
Market Research
Combine structured data with text to evaluate changing business conditions.
Financial Operations
Improve reporting, anomaly detection, resource planning, and decision workflows.
Investment Screening
Rank opportunities using transparent, repeatable analytical criteria.
Scenario Analysis
Test how financial outcomes change under different assumptions and shocks.
Financial collaboration
Better financial decisions begin with better evidence.
Machine learning can help organizations organize information, test assumptions, evaluate risk, and identify patterns that may be difficult to see through manual analysis alone.