KoeusIss

Keybase proof

I hereby claim:

• I am koeusiss on github.
• I am koeusiss (https://keybase.io/koeusiss) on keybase.
• I have a public key ASCSJc_YSGbCLa4WegQu2KJ3w8XfZ4-q4xvg3RM182dQpQo

To claim this, I am signing this object:

KoeusIss

LinearRegression: RMSPE 0.30116
ElasticNet: RMSPE 0.31334
DecisionTreeRegressor: RMSPE 0.29511
MLPRegressor: RMSPE 0.54968
LGBMRegressor: RMSPE 0.29734
BaggingRegressor: RMSPE 0.29085
RandomForestRegressor: RMSPE 0.29252
ExtraTreesRegressor: RMSPE 0.29104
Super Learner: RMSPE 0.26938

KoeusIss

## All prediction algorithm wrappers in SuperLearner:
## -----------------------------------------------------------------------
##  [1] "SL.bartMachine"      "SL.bayesglm"         "SL.biglasso"        
##  [4] "SL.caret"            "SL.caret.rpart"      "SL.cforest"         
##  [7] "SL.earth"            "SL.extraTrees"       "SL.gam"             
## [10] "SL.gbm"              "SL.glm"              "SL.glm.interaction" 
## [13] "SL.glmnet"           "SL.ipredbagg"        "SL.kernelKnn"       
## [16] "SL.knn"              "SL.ksvm"             "SL.lda"             
## [19] "SL.leekasso"         "SL.lm"               "SL.loess"           
## [22] "SL.logreg"           "SL.mean"             "SL.nnet"            

KoeusIss

def conv_backward(dZ, A_prev, W, b, padding="same", stride=(1, 1)):
    """
    Performs back-propagation over a convolutional layer of a neural network
    """
    m, h_new, w_new, c_new = dZ.shape
    m, h_prev, w_prev, c_prev = A_prev.shape
    kh, kw, c_prev, c_new = W.shape
    sh, sw = stride

    dA_prev = np.zeros(A_prev.shape)

KoeusIss

# importing
import tensorflow as tf
# alpha (float): Is the original learning rate
# decay_rate (float): Is the weight used to determine the rate at which alpha will decay.
# global_step (int): Is the number of passes of gradient descent that have elapsed.
# decay_step (int): Is the number of passes of gradient descent that occur before alpha is decayed further.
tf.train.inverse_time_decay(
  alpha, global_step, decay_step, decay_rate, staircase=True
)

KoeusIss

# importing
import tensorflow as tf
# alpha: is the learning rate
# beta1: is the first moment weight
# beta2: is the second moment weight
adam = tf.train.AdamOptimizer(alpha, beta1, beta2, epsilon)
adam.minimize(loss)

KoeusIss

# importing
import tensorflow as tf
# alpha: the learning rate
# beta: rmsprop weight
rms_prop = tf.train.RMSPropOptimizer(alpha, beta, epsilon=epsilon)
rms_prop.minimize(loss)

KoeusIss

# Import
import numpy as np
import matplolib.pyplot as plt
# Define the input x axis value
x = np.arange(-1, 1, 0.01)
# Define the Heaviside function suing numpy builtin np.heaviside(x, y) where
# x is the input (x axis) and y the value of the function when x is 0.
y = np.heaviside(x, 0)
plt.plot(x, y)

KoeusIss

>>> def demo(x):
...    print("x: {} -- id: {}".format(x, id(x)))
...    x = 98
...    print("x: {} -- id: {}".format(x, id(x)))
...
>>> demo(1)
x:1 -- id: 93973708342016
x:98 -- id: 93973708345120

KoeusIss

>>> tpl = ([1, 2], 3)
>>> tpl
([1, 2], 3)
>>> id(tpl)
139678802945544
>>> tpl[0][0] = "a"
>>> tpl
(['a', 2], 3)
>>> id(tpl)
139678802945544