From: Temporal topic model for clinical pathway mining from electronic medical records
Symbol | Description |
---|---|
\(\mathcal {A}\) | The set of all treatment activities; |
a | The treatment activity; |
\(\mathcal {D}\) | The set of all treatment days; |
d | The treatment day; |
K | The number of treatment topics; |
\(N_{d}\) | The number of all treatment activities in a treatment day; |
\(\theta\) | The probability distribution of topic in treatment day; |
\(\varphi\) | The probability distribution of timestamp in treatment topic; |
\(\phi\) | The probability distribution of activity in treatment topic; |
\(\alpha , \delta , \beta\) | The hyper-parameters; |
z | The treatment topic; |
t | The treatment timestamp of treatment day; |
T | The universe of treatment days; |
\(\mathcal {L}\) | The whole treatment processes; |
\(Z_{N}\) | The normalization factor; |
N | The top activity number in the ranking results; |
\(n_{d,k}\) | The number of times that the day d is assigned to topic k; |
\(q_{k,t}\) | The number of times that the timestamp t is assigned to topic k; |
\(m_{k,t,a}\) | The number of times that the activity a is assigned to topic k with timestamp t; |
\(\sigma\) | A treatment process for a specific patient; |
\(\mathcal {T}\) | The set of treatment timestamp; |
C | The triples form of discovered CP; |