Skip to main content
Have a personal or library account? Click to login
Personalized Recommendation Multi-Objective Optimization Model Based on Deep Learning Cover

Personalized Recommendation Multi-Objective Optimization Model Based on Deep Learning

By: Zepeng Yang,  Ping Lu and  Pingping Liu  
Open Access
|Mar 2024

Figures & Tables

Figure 1.

Model structure of the Wide & Deep Learning model.

Figure 2.

Shared-Bottom model.

Figure 3.

FM_Shared Expert Multi-Objective Dependency (FSMD) Mode

Figure 4.

Multi-objective Base Model.

Figure 5.

Multi-gate MoE Model.

Figure 6.

Internal Structure of Gated Network.

Figure 7.

FM_Gate (FG) Model.

Figure 8.

FM_Sharing the deep part of the Expert (FS) Model.

Figure 9.

FM_Shared Expert Multi-Objective Dependency(FSD)Model.

Figure 10.

The seesaw phenomenon in each model under complex target association

Figure 11.

AUC comparison of network shapes.

Figure 12.

AUC comparison of the number of layers in a gated network.

Figure 13.

AUC comparison of the number of experts.

PARAMETER SETTINGS

Parameter NameValue
batch_size256
optimizeradam
learning_rate0.001
embedding_size4
dnn_layers(512, 256)
dnn_use_bnTrue

Results of the UCI Census Income Dataset

ModelsAUC/ IncomeAUC/ MaritalMean
Single-Task0.91980.97480.9473
Shared-Bottom0.91480.97540.9451
MMoE0.91520.97560.9454
PLE0.91610.97640.9463
Base0.91340.97060.9420
FG0.91460.96930.9420
FS0.92160.97560.9486

Parameter Settings

ModelsAUC/ IncomeAUC/ MaritalMean
Single-Task0.71920.66350.6914
DeepFM0.7190.6580.689
Shared-Bottom0.71840.69160.705
ESMM0.71890.68970.7043
MMoE0.72010.70570.7129
Base0.71930.70860.714
FG0.72010.70840.7143
FS0.72040.71140.7159
FSMD0.72050.71170.7161
Language: English
Page range: 44 - 57
Published on: Mar 28, 2024
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2024 Zepeng Yang, Ping Lu, Pingping Liu, published by Xi’an Technological University
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.