AllenYGY's Blog

Appearance

Causality Abstract ​

What we want to do ​

Construct a casual graphical model to predict movie box office

What is Causal Graphical Model ​

Construct Causal Graph manually depend on past experience ​

Construct Causal Graph automatically ​

Motivation ​

  1. Understand Interactions Between Variables
  2. Discovering Potential Causal Relationships
  3. Data-Driven Experimentation

Understand Interactions Between Variables ​

Suppose Variables Independent

  1. Budget: The amount spent on the production of the movie directly impacts the quality of production, marketing, and distribution.
  2. Genres: Different genres attract different audiences and can affect box office performance depending on trends and audience preferences.
  3. Cast: The actors in a movie can draw in audiences, especially if the cast includes popular or well-known stars.
  4. Release Date: Timing plays a critical role, as certain periods (e.g., summer, holidays) typically see higher ticket sales.

Mining potential causal relationships between variables ​

A B 存在因果性， 但不存在相关性

Get counter-intuitive Conclusions ​

Simpson's Paradox

Data-Driven Experimentation ​

Multi-linear Regression

Time-Series Analysis

Hybrid model Analysis

References ​

  1. Wooldridge, J. M. (2016). Introductory econometrics: A modern approach (6th ed.). Cengage Learning.

  2. Weisberg, S. (2005). Applied linear regression (3rd ed.). Wiley.

  3. Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time series analysis: Forecasting and control (5th ed.). Wiley.

  4. Hamilton, J. D. (1994). Time series analysis. Princeton University Press.

  5. Tsai, C. F., & Hung, C. S. (2014). Modeling credit scoring using neural network ensembles. Neural Computing and Applications, 24(2), 445–452. https://doi.org/10.1007/s00521-012-1247-4

  6. Zhang, G. P. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50, 159-175. https://doi.org/10.1016/S0925-2312(01)00702-0