Table 2.
A summary of notations used in WJST.
| Symbol | Description |
|---|---|
| Collections | |
| R | Set of all documents |
| V | Vocabulary set |
| Q | Set of all distinct windows (with different sizes) |
| Z | Set of all topics |
| E | Set of all distinct weights |
| S | Set of all sentiment labels |
|
| |
| Init parameters | |
| q | Window variable |
| e | Weight variable |
| r | Document variable |
| z | Topic variable |
| w | Word variable |
| s | Sentiment variable |
|
| |
| Distributions | |
| θ | Probability of z given s and r |
| φ | Probability of w given z, s, q, and e |
| π | Probability of s given r (s0 = Positive label, s1 = Negative label) |
| ψ | Probability of e given z, s, and r |
| ξ | Probability of q given z and r |
|
| |
| Hyper parameters | |
| α | Dirichlet prior distribution for θ |
| β | Dirichlet prior distribution for φ |
| γ | Dirichlet prior distribution for π |
| δ | Dirichlet prior distribution for ψ |
| μ | Dirichlet prior distribution for ξ |