Radial Basis Function Neural Network Based on an Improved Exponential Decreasing Inertia Weight-Particle Swarm Optimization Algorithm for AQI Prediction

Table 2

The parameters of RBFNN by EDIW-PSO: centers 𝑐_𝑗𝑖, widths 𝑑_𝑗, and connection weights 𝜔_𝑗𝑘.

𝑗	1	2	3	4	5	6	7	8	9	10
𝑐𝑗1	1.0000	−0.9849	−0.9569	0.7735	0.5874	1.0000	−0.4653	−1.0000	0.7688	−0.3863
𝑐𝑗2	0.5752	−0.1652	−0.9023	−0.9643	0.6898	0.4439	0.4378	0.4779	0.0860	−0.8635
𝑐𝑗3	0.4994	0.2723	−1.0000	−0.1271	0.7500	0.5629	0.2198	−0.9292	0.7286	−0.9566
𝑐𝑗4	0.5250	−0.2811	0.7298	0.9254	−0.1321	0.3404	−0.3397	−0.1867	0.5373	−0.6491
𝑐𝑗5	−0.0776	−1.0000	0.1190	0.4846	0.2787	0.1364	−0.8673	0.9034	1.0000	0.0959
𝑐𝑗6	−0.4532	0.9372	1.0000	0.5370	0.2975	0.9214	0.4095	−1.0000	−0.0813	−0.2635
𝑐𝑗7	−0.0124	1.0000	−0.4612	0.0190	−0.1061	−0.2451	0.7454	0.1273	0.1925	0.0609
𝑑𝑗	−0.0156	−0.9724	0.0728	0.0588	−0.5133	0.8992	1.0000	−0.0857	0.7017	−0.9959
𝜔𝑗	−0.0289	−0.7936	−0.5968	0.8891	−0.5265	−0.0812	0.3979	0.2548	−0.9623	0.8871