This is the public repo for CoEA: A Cooperative-Competitive Evolutionary Algorithm for Bidirectional Recommendations
GT is our Algorithm(GTEA(-GM);GTEA(-GC);GTEA) ! Note that dataset is too large, we don't include it in this repo, please leave your email if you want.
Algorithm Step
-
PlatEMO-RUN Benchmarks(NSGA-Ⅱ、SPEA2、MOEA-D and ARMOEA etc.)
Run main.m, parameter settings as follows:
Results will be saved in data directory, then run Results.m to get correct format (PS/PSV.mat).
Next, evalute the results
-
First, run Repair_MOEAs.m, get PS_Repair.mat;
-
Second, run Evaluate_MOEAs.m, get PSV_Repair.mat;
-
Third, run Non_dominate.m, get PS_Nondo.mat and PSV_Nondo.mat
-
-
P2P_Lending problem
-
First, Run UCF and ICF to get the corr result.( run CFCommendation.m, get UCF.mat; then run ICF Commendation.m, get ICF.)
ICF should be run twice(modify the corr UCF or ICF)
Note: ps =cell(1,1)ps{1}= PS (ICF)
-
Second, run WholeFramework.m
-
Third, run draw.m, get Pareto Frontier(https://jingyan.baidu.com/article/6fb756ec6dc922241858fb19.html)(http://www.pudn.com/Download/item/id/2148827.html)。
-
Finally, run Matric.m, get the evalution metrics.
-
-
Group_Purchase Problem
- The same as P2Plending.
-
Note: Matlab Storage format must be modified in V7.3mat
If you follow our work, please cite it:
[1] Zhao H, Wu X, Zhao C, et al. CoEA: A Cooperative-Competitive Evolutionary Algorithm for Bidirectional Recommendations[J]. IEEE Transactions on Evolutionary Computation, 2021.