Time-Aware Explainable Recommendation via Updating Enabled Online Prediction

. 2022 Nov 11;24(11):1639. doi: 10.3390/e24111639

Algorithm 1: Replacing-updating Enabled Online Recommendation

Require:
Initial training set $D_{0}$ , updating cycle T, new received data during updating cycle $D_{T}$
Ensure:
Recommender system $R S$ , Recommendations with explanations: $R e s u l t s$
1:
//Initial training phase
2:
$D \leftarrow D_{0}$
3:
$R S \leftarrow t r a i n (D)$
4:
//Updating phase
5:
for each cycle T do
6:
$D \leftarrow D_{T}$
7:
$R S \leftarrow t r a i n (D)$
8:
end for
9:
//Inference phase
10:
for each user u do
11:
$R e s u l t s \leftarrow r e c o m m e n d (R S, u)$
12:
end for