Enhancing energy consumption prediction in smart homes: a convergence-aware federated transfer learning approach | Science and Technology for Energy Transition (STET)

Emerging Advances in Hybrid Renewable Energy Systems and Integration

Open Access

Table 5

Comparison of proposed FedDNN and FedLSTM models with existing models in terms of R2 score.

Algorithm	C1	C2	C3	C4	C5	C6	C7	C8	Average
LinearRegression	0.17	0.23	0.155	0.05	0.19	0.14	0.087	0.20	0.15
LGBMRegressor	80.04	84.76	85.05	87.93	89.19	86.05	81.72	85.04	84.97
XGBRegressor	80.11	84.80	85.09	88.00	89.25	86.10	81.77	85.10	85.03
CatBoostRegressor	80.08	84.78	85.07	87.96	89.23	86.08	81.75	85.08	85.00
SGDRegressor	0.16	0.22	0.15	0.03	0.18	0.14	0.07	0.20	0.14
KernelRidge	−6.42	−5.06	−5.49	−7.04	−5.87	−5.98	−6.91	−6.36	−6.14
BayesianRidge	0.17	0.23	0.15	0.04	0.19	0.14	0.08	0.20	0.15
GradientBoostingRegressor	74.22	80.97	79.98	83.18	84.34	80.52	76.50	79.38	79.89
SVR	0.94	0.49	0.16	0.61	0.79	0.44	0.45	1.01	0.61

DNN	82.02	83.00	84.25	88.27	88.68	85.08	82.02	85.04	84.80
FedDNN-full	82.99	86.41	86.39	89.98	89.25	85.43	83.47	86.86	86.35
FedDNN-par	83.93	89.42	86.39	91.87	92.93	85.65	84.24	88.34	87.85

LSTM	76.34	85.23	84.91	88.57	88.94	87.87	83.03	86.13	85.13
FedLSTM-full	79.04	86.95	84.55	90.25	87.83	88.38	84.11	86.69	85.98
FedLSTM-par	84.33	89.31	87.27	90.46	92.06	90.18	85.73	90.13	88.68

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.