Fig 1.
HSQ index framework.
Fig 2.
The mechanism of happiness.
Fig 3.
Structure of a Multilayer Perceptron.
Fig 4.
The IGSA-MLPNN-GARSON framework.
Fig 5.
Study area.
Fig 6.
Spatial and age distribution of valid respondents.
(a) District-level sample distribution. (b) Age distribution of respondents.
Fig 7.
(a) X11-X44, (b) X45-X84.
Fig 8.
Residents’ happiness scores.
Fig 9.
Tuning process of the IGSA-MLPNN model.
(a) adjustment of the number of hidden neurons and its impact on the MAE; (b) parallel coordinate visualization.
Fig 10.
(a) Primary indicator categories. (b) Ecological habitability. (c) Diversity and inclusiveness. (d) Safety toughness. (e) Convenient transportation. (f) Cityscape features. (g) Healthy and comfortable. (h) Neat and orderly. (i) Innovation vitality.
Table 1.
Accuracy comparison based on 10-fold cross-validation (mean ± SD).
Fig 11.
Accuracy comparison of different models in terms of the MAE, MSE, RMSLE, and R² based on 10-fold cross-validation.
The error bars denote 95% confidence intervals.