Features: P/E Cash Flow total Debt Industry Returns Equally weighted industry return value weighted industry return Output variable: How well the industry is perofmring Measure: the relationship between the ratio corroreraltion with indsutry performance Feature: Ablative, Additive Baseline: Lin. Reg Valuation ratios book values Standard ones Probabability ratio Industry Returns: Overall economy is doing - S&P Number of companies that remain in the industry
Split set: 2 Linear regression: 1 linear regression models: features background reading: 1 paragraph on ratios cite existing references
Google Docs: Dump links: Motivation: What problem are you tackling, and what's the setting you're considering?
Method: What machine learning techniques have you tried and why? (Linear Regression 2 baseline models) Preliminary experiments: Describe the experiments that you've run, the outcomes, and any error analysis that you've done. You should have tried at least one baseline. (Error analysis MSE)
\To:do Timeseries forecast
Data split: val split dev splits-> 2009-2017
DataseT: https://l.facebook.com/l.php?u=https%3A%2F%2Fwww.gsb.stanford.edu%2Flibrary%2Farticles%2Fdatabases%2Flinks%2Ffinancial-ratios-suite%3Ffbclid%3DIwAR2Yao0Vlwo-UfM6015X5YdqG5pRoWTo-wOLpjxkYjWMXZU7vzxtKx3bivk&h=AT2-dDt_a_dKkG_NT31kLolPaXnuJJmGKDPsmN74DNlqT-iXAt8_F3LkuVylm5_6zJdEvOPX6-XjSWIlxkbXTw0SjYDIOl31lQRB_1-UnKL60a5L2o40flV0gxxddz1P3eZF7e65Aqw
Validation split: by Month Control industries: run and experiment:
@czhang
@kinbert
Linear Regression Model: time series
Linear Regression Model: carolines idea
hopefully implemented by tomorrow evening.
Next steps: Given your preliminary results, what are the next steps that you're considering? (Sunday)
Dev set: Ratio datasets industry returns: output value: -> starting with few features supervised learning: if they are good classification: adding more features into it
Datasets:
'ee6d2f60cdafb550.csv' -- Fama-French 49 Industries, 74 Variables 76 Features, YYMMDDs10 Date