Tiêu đề 8.2. Background.pdf ✅

pacmann.io © 2022 – Pacmann AI 5 Background: Estimating Causal Effect
pacmann.io © 2022 – Pacmann AI 6 Backdoor Path ● We can identify causal relationships by acknowledging and blocking the influence of backdoor paths: ○ Model the data generating process using causal diagram ○ List all paths ○ Determining back doors or alternate explanations for why a relationship might occur. ○ Closing down back door paths e.g by controlling variables
pacmann.io © 2022 – Pacmann AI 7 Backdoor Path Example we want to know the effect of IPspend on profit: 1. Model the data generating process using causal diagram
pacmann.io © 2022 – Pacmann AI 8 Backdoor Path Example we want to know the effect of IPspend on profit: 2. List all paths ● IPspend → profit ● IPspend ← tech→ profit