如何制作动态回归模型

Question

我有2个数据集，分别是FPDataset和SIDataset。然后是 FPDatset 和 SIDataset 然后每个数据集将被分成两组。 FPDataset：FPWalking 和 FPStanding SIDataset：SIWalking 和 SIStanding

我的数据集的插图如下：

FPWalking = np.array([
    [0.0, 0.0, 0.0, 0.0],
    [0.0, 0.0, 0.0, 0.0],
    [5.29, -1.22, 8.282, 8.22],
    [-1.92, 9.781, 3.28, 1.93],
    [0.0, 0.0, 0.0, 0.0],
    [0.0, 0.0, 0.0, 0.0],
    [0.27, 6.12, 8.12, 2.34],
    [6.93, 2.83, 0.72, 7.22]
])

FPStanding = np.array([
    [6.93, 2.83, 0.72, 7.22],
    [0.27, 6.12, 8.12, 2.34],
    [5.29, -1.22, 8.282, 8.22],
    [-1.92, 9.781, 3.28, 1.93],
])

SIWalking = np.array([
    [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    [9.0, 92.0, 0.0, 0.0, 0.0, 1.0, 21.0, 71.0, 0.0, 12.0, 65.0, 45.0],
    [9.0, 92.0, 0.0, 0.0, 0.0, 1.0, 21.0, 71.0, 0.0, 12.0, 65.0, 45.0],
    [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    [0.0, 0.0, 45.0, 0.0, 0.0, 22.0, 45.0, 23.0, 0.0, 77.0, 89.0, 0.0],
    [0.0, 0.0, 45.0, 0.0, 0.0, 22.0, 45.0, 23.0, 0.0, 77.0, 89.0, 0.0],
])

SIStanding= np.array([
    [0.0, 0.0, 0.0, 0.0, 73.0, 12.0, 66.0, 78.0, 0.0, 12.0, 87.0, 0.0],
    [0.0, 0.0, 0.0, 0.0, 73.0, 12.0, 66.0, 78.0, 0.0, 12.0, 87.0, 0.0],
    [0.0, 0.0, 0.0, 0.0, 73.0, 12.0, 66.0, 78.0, 0.0, 12.0, 87.0, 0.0],
    [0.0, 0.0, 0.0, 0.0, 73.0, 12.0, 66.0, 78.0, 0.0, 12.0, 87.0, 0.0],

我想使用 SIDataset（SIWalking 和 SIStanding）预测 FPDataset（FPWalking 和 FPStanding）。 现在我已经制作了一个模型，然后只需连接数据集如下：

SIDataset = np.concatenate((SIDatasetWalking,SIDatasetStanding), axis=0)
FPDataset = np.concatenate((FPDatasetWalking,FPDatasetStanding), axis=0)

在我连接数据集之后，我训练了一个模型，使用 resnet 来解决我的问题。但我的模型只适用于步行数据。

我的架构代码：

def identity_block(input_tensor, filters):
    x = layers.Conv1D(filters, kernel_size=5, padding='same')(input_tensor)
    x = layers.Activation('relu')(x)

    x = layers.Conv1D(filters, kernel_size=5, padding='same')(x)
    x = layers.Activation('relu')(x)

    x = layers.Conv1D(filters, kernel_size=5, padding='same')(x)

    x = layers.add([x, input_tensor])
    x = layers.Activation('relu')(x)

    return x

def dens_block(input_tensor, filters):
    x = layers.Conv1D(filters, kernel_size=5, padding='same')(input_tensor)
    x = layers.Activation('relu')(x)

    x = layers.Conv1D(filters, kernel_size=5, padding='same')(x)
    x = layers.Activation('relu')(x)

    x = layers.Conv1D(filters, kernel_size=5, padding='same')(x)

    shortcut = layers.Conv1D(filters, kernel_size=5, padding='same')(input_tensor)

    x = layers.add([x, shortcut])
    x = layers.Activation('relu')(x)
    return x

def ResNet50Regression():
    input_shape = (89, 1)  # Change the number of input features (channels) accordingly
    width = 32

    Res_input = layers.Input(shape=input_shape)

    x = dens_block(Res_input, width)
    x = identity_block(x, width)
    x = identity_block(x, width)

    x = dens_block(x, width*2)
    x = identity_block(x, width*2)
    x = identity_block(x, width*2)

    x = dens_block(x, width*4)
    x = identity_block(x, width*4)
    x = identity_block(x, width*4)

    x = layers.GlobalAveragePooling1D()(x)
    x = layers.Dense(6, activation='linear')(x)
    model = models.Model(inputs=Res_input, outputs=x)

    return model

训练结果预测：

下面是步行预测结果

以下为常设预测结果

如何让我的模型可以预测站立和步行数据？因为现在该模型仅适用于步行数据

我的完整代码在这里：https://colab.research.google.com/drive/1ajARJWJrYY3Qa7XKldJirX6QHutQzkzc?usp=sharing