I have recently begun coding deep neural networks in Python and I have been stuck with this problem for weeks. I have checked other similar questions but could not grasp the solution.
I have a feed forward neural network and I am trying to obtain the R^2 value for my model. I have a dataframe of 1031 rows by 9 columns. I am fitting the first 7 columns (storm characteristics) with the water elevation as given in 8th or 9th column. Here 'h' is column heading that I am fitting with the first 7 columns.
See csv data at this link.
### Load dependencies
import pandas as pd
import numpy as np
from sklearn.neural_network import MLPRegressor
from sklearn.metrics import accuracy_score, mean_squared_error
from sklearn.model_selection import cross_val_score
from sklearn.metrics import r2_score
data = pd.read_csv (r'C:\Users\SM\Desktop\dataframe.csv')
### Define MLP Hyper-parameters
n = 20 # ----------- Number of layers
p = 450 # ---------- Number of perceptrons in each layer
i = 1 # ------------ Initializing iteration count
MLPsize = []
### While loop to create tuple for MLP hyperparameter input
while i <= n:
MLPsize.append(p)
i += 1
Model = MLPRegressor(hidden_layer_sizes=tuple(MLPsize),
activation='logistic', solver='adam', alpha=0.0001,
batch_size='auto', learning_rate='adaptive', learning_rate_init=0.001,
power_t=0.5, max_iter=5000, shuffle=False, random_state=None, tol=0.0001,
verbose=False, warm_start=False, momentum=0.9, nesterovs_momentum=True,
early_stopping=False, validation_fraction=0.5,
beta_1=0.9, beta_2=0.999, epsilon=1e-08, n_iter_no_change=10)
### Define function to run train-test simulation
msk = np.random.rand(len(data)) < 0.7
Train = data[msk]
Test = data[~msk]
Train = Train.reset_index(drop=True)
Test = Test.reset_index(drop=True)
def RunModel (h):
global Train_pred
global Test_pred
global Train_true
global Test_true
global Train_error
global Test_error
Train_error = []
Test_error = []
Model.fit(Train.iloc[:,:7], Train[h])
Train_pred = Model.predict(Train.iloc[:,:7])
Train_error.append(mean_squared_error(Train[h], Train_pred))
Train_true = np.array(Train[h]).reshape(-1, 1)
Train_pred = Train_pred.reshape(-1, 1)
Test_pred = Model.predict(Test.iloc[:,:7])
Test_error.append(mean_squared_error(Test[h], Test_pred))
Test_true = np.array(Test[h]).reshape(-1, 1)
Test_pred = Test_pred.reshape(-1, 1)
print(mean_squared_error(Train_pred, Train_true))
print(Model.score(Test_pred, Test_true))
RunModel('11939')
And I get an error:
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-23-01acc869ac62> in <module>
----> 1 Model.score(Test_pred, Test_true)
C:\ProgramData\Anaconda3\lib\site-packages\sklearn\base.py in score(self, X, y, sample_weight)
406 from .metrics import r2_score
407 from .metrics.regression import _check_reg_targets
--> 408 y_pred = self.predict(X)
409 # XXX: Remove the check in 0.23
410 y_type, _, _, _ = _check_reg_targets(y, y_pred, None)
C:\ProgramData\Anaconda3\lib\site-packages\sklearn\utils\extmath.py in safe_sparse_dot(a, b, dense_output)
140 return ret
141 else:
--> 142 return np.dot(a, b)
143
144
**ValueError: shapes (295,1) and (7,450) not aligned: 1 (dim 1) != 7 (dim 0)**
Please help. Any detailed explanation would be appreciated.
The line that raises the error is the following:
print(Model.score(Test_pred, Test_true))
This is not valid since .score() method of the MLPRegressor takes as inputs (X,y). Read more here: https://scikit-learn.org/stable/modules/generated/sklearn.neural_network.MLPRegressor.html#sklearn.neural_network.MLPRegressor.score
So, if you want to estimate the R2 coefficient of the model you need to pass the Xtest and ytest that in your case are Test.iloc[:,:7], Test_true.
### Load dependencies
import pandas as pd
import numpy as np
from sklearn.neural_network import MLPRegressor
from sklearn.metrics import accuracy_score, mean_squared_error
from sklearn.model_selection import cross_val_score
from sklearn.metrics import r2_score
data = pd.read_csv (r'C:\Users\SM\Desktop\dataframe.csv')
### Define MLP Hyper-parameters
n = 20 # ----------- Number of layers
p = 450 # ---------- Number of perceptrons in each layer
i = 1 # ------------ Initializing iteration count
MLPsize = []
### While loop to create tuple for MLP hyperparameter input
while i <= n:
MLPsize.append(p)
i += 1
Model = MLPRegressor(hidden_layer_sizes=tuple(MLPsize),
activation='logistic', solver='adam', alpha=0.0001,
batch_size='auto', learning_rate='adaptive', learning_rate_init=0.001,
power_t=0.5, max_iter=5000, shuffle=False, random_state=None, tol=0.0001,
verbose=False, warm_start=False, momentum=0.9, nesterovs_momentum=True,
early_stopping=False, validation_fraction=0.5,
beta_1=0.9, beta_2=0.999, epsilon=1e-08, n_iter_no_change=10)
### Define function to run train-test simulation
msk = np.random.rand(len(data)) < 0.7
Train = data[msk]
Test = data[~msk]
Train = Train.reset_index(drop=True)
Test = Test.reset_index(drop=True)
def RunModel (h):
global Train_pred
global Test_pred
global Train_true
global Test_true
global Train_error
global Test_error
Train_error = []
Test_error = []
Model.fit(Train.iloc[:,:7], Train[h])
Train_pred = Model.predict(Train.iloc[:,:7])
Train_error.append(mean_squared_error(Train[h], Train_pred))
Train_true = np.array(Train[h]).reshape(-1, 1)
Train_pred = Train_pred.reshape(-1, 1)
Test_pred = Model.predict(Test.iloc[:,:7])
Test_error.append(mean_squared_error(Test[h], Test_pred))
Test_true = np.array(Test[h]).reshape(-1, 1)
Test_pred = Test_pred.reshape(-1, 1)
print(mean_squared_error(Train_pred, Train_true))
print(Model.score(Test.iloc[:,:7], Test_true))
RunModel('11939')
Prints:
0.5131277608869395
-0.02165748764016695