各位大侠，帮帮忙怎么才能得到偏最小二乘法的回归方程，最好有详细步骤。

数据如下，谢谢了。

x1 x2 x3 x4 x5 x6 x7 y
567.21 36574 1660 304923 535 83.0 18.0 6235
573.89 44719 1540 311590 542 82.0 19.7 6897
581.21 55828 1594 340400 585 85.0 20.5 7012
585.75 50629 1499 407773 693 85.5 21.0 7023
588.89 57645 1600 470404 795 86.0 21.3 7289
592.43 79552 1947 572569 962 86.8 22.7 7896
598.72 105800 1950 660180 1100 87.0 24.0 7589
607.21 150016 1955 847263 1394 84.0 23.8 7986
615.90 182673 1966 1150970 1867 98.0 24.1 7998
619.91 123278 1999 1258556 2010 97.0 24.2 8012
625.27 159147 2430 1485282 2362 99.0 25.4 8123
626.27 208114 2465 1721637 2720 100.0 26.2 8456
629.00 414659 2600 2296046 3624 110.0 26.8 8498
629.93 595437 2610 3254235 5043 140.0 28.3 8654
631.63 711650 2732 4278600 6730 273.0 54.3 8723
634.88 696761 2700 5394000 8451 287.2 90.3 8923
638.37 661720 2530 6152400 9589 132.8 96.8 10093
641.48 683437 2200 6750000 10466 128.0 107.3 11626
643.35 973217 2230 7400000 11439 137.1 109.3 11536
644.79 983256 2210 8006600 12345 135.2 106.9 11276
645.80 995612 2256 8795900 13546 108.3 111.8 11309

用优化软件的结论：
相关系数(R): 0.98113306532849
相关系数之平方(R^2): 0.962622091880879
决定系数(DC): 0.962404522931815
卡方系数(Chi-Square): 117.089904659981
F统计(F-Statistic): 489.32165190862

参数 最佳估算
---------- -------------
a1 12.1037143195345
a2 0.0011184295022518
a3 -2.68867427023649E-7
a4 2.73455242044292
a5 0.544960363781314
a6 -3.40429595027621
a7 10.6753437587823

====== 结果输出 =====

No 实测值y 计算值y
1 6235 6971.2314627
2 6897 7093.6367462
3 7012 7199.5423429
4 7023 7292.6927612
5 7289 7357.5984765
6 7896 7430.7809032
7 7589 7578.2523295
8 7986 7799.4861779
9 7998 7991.3757428
10 8012 7987.8188120
11 8123 7996.6038224
12 8456 8093.0186609
13 8498 8357.8857016
14 8654 8579.6781854
15 8723 8614.9029693
16 8923 9138.2322798
17 10093 10086.9135086
18 11626 10947.4402900
19 11536 11370.4470120
20 11276 11552.8750770
21 11309 11799.3025397