Macroeconomic Forecasting Using Alternative Data: Techniques for Applying Big Data and Machine Learning applies computer science to the demands of macroeconomic forecasting. It is the first book to combine machine learning methods with macroeconomics. By using artificial intelligence and machine learning techniques, it unlocks the increased forecasting accuracy offered by alternative data sources. Through its interdisciplinary approach, readers learn how to use big datasets efficiently and effectively.
- Combines big data/machine learning with macroeconomic forecasting
- Explains how alternative data improves forecasting accuracy when controlled for traditional data sources
- Provides new innovative methods for handling large databases and improving forecasting accuracy
2. Macro Data are Noisy
3. Our Goal: Macro Data with Less Noise and Lag
4. Alternate Data
5. A Framework for Alternate Data
6. Predicting Data Releases with Search
7. Modeling Case Study: Non-Farm Payrolls
8. Accounting Data
9. Prediction in Practice
10. Public Good: Visualizing World Economic Growth in Real Time
11. Interviews with Policy Makers and Asset Managers
Apurv Jain is the Senior Finance Lead and Co-Founder of the Economic Measurement Group at Microsoft. His team of scientists from Microsoft Research, ML experts from BingPredicts, and traders from Capital Markets Group use web-scale data (search, twitter etc.) to understand and predict the economy and the financial markets. Apurv sets the external product and research agenda, and he is the portfolio manager for an internal $150 mm portfolio devoted to testing our ideas. His alternate data and AI based strategies have a positive 3 year track record. He is also a visiting researcher at Harvard Business School.