核转换函数 Args: X dataMatIn数据集 A dataMatIn数据集的第i行的数据 kTup 核函数的信息 Returns:
(X, A, kTup)
| 48 | |
| 49 | |
| 50 | def kernelTrans(X, A, kTup): # calc the kernel or transform data to a higher dimensional space |
| 51 | """ |
| 52 | 核转换函数 |
| 53 | Args: |
| 54 | X dataMatIn数据集 |
| 55 | A dataMatIn数据集的第i行的数据 |
| 56 | kTup 核函数的信息 |
| 57 | |
| 58 | Returns: |
| 59 | |
| 60 | """ |
| 61 | m, n = shape(X) |
| 62 | K = mat(zeros((m, 1))) |
| 63 | if kTup[0] == 'lin': |
| 64 | # linear kernel: m*n * n*1 = m*1 |
| 65 | K = X * A.T |
| 66 | elif kTup[0] == 'rbf': |
| 67 | for j in range(m): |
| 68 | deltaRow = X[j, :] - A |
| 69 | K[j] = deltaRow * deltaRow.T |
| 70 | # 径向基函数的高斯版本 |
| 71 | K = exp(K / (-1 * kTup[1] ** 2)) # divide in NumPy is element-wise not matrix like Matlab |
| 72 | else: |
| 73 | raise NameError('Houston We Have a Problem -- That Kernel is not recognized') |
| 74 | return K |
| 75 | |
| 76 | |
| 77 | def loadDataSet(fileName): |
no outgoing calls
no test coverage detected