Algorithm 1: LDA's generative process

→ For each document $d=1\text{,}$ $\dots \text{,}M$ in the corpus $C$: 1. Sample mixing probability ${\mathrm{\theta }}_{\mathrm{d}}^{\sim }\text{Dir}\left(\mathrm{\alpha }\right)$ 2. For each of $\mathrm{V}$ words ${\mathrm{w}}_{\text{di}}$: a) Select a topic ${\mathrm{z}}_{\text{di}}\in {\left\{1,\dots ,\mathrm{K}\right\}}^{\sim }$ Multinomial (${\mathrm{\theta }}_{\mathrm{d}})$ b) Select a word ${\mathrm{w}}_{\text{di}}\in {\left\{1,..,\mathrm{V}\right\}}^{\sim }$ Multinomial (${\mathrm{\beta }}_{{\mathrm{z}}_{\text{di}}})$