cheatsheet.txt

##################################################################################################
#                                                                                                #
#    collection of some useful pandas,Numpy and matplotlib function useful in times of data wrangling and processing.       #                                                                                                     #                                                                                                #
##################################################################################################


import pandas as pd

---make data frame---
df = pd.read_csv('train.csv',delimiter="\t",header=0,names=[])

df.dtypes

df.head() --> print top 5 data

df.columns.values ---> all columns names as list

df["column_name"].median()

df['colum_name'] -->>dtype of series
df[["co1","col2","col3"]]

df["new column"] = value
df.info()

df.describe()

df.describe()["column_name"] or df["column_name"].describe()

df['column_name'] = df.['column_name'].fillna(value) --> to replace empty or nan

df.loc[df["column_name"]=="male","column_name"]=0 -->to locate and replace male with o i nsex column

df["column_name"].unique() -->gives all the unique values in embarked column

df[df['Age'] > 60]
df[df['Age'] > 60][['Sex', 'Pclass', 'Age', 'Survived']]
len(df[ (df['Sex'] == 'male') & (df['Pclass'] == i) ])
df['Gender'] = df['Sex'].map( {'female': 0, 'male': 1} ).astype(int)

df.drop([col1,col2,....],axis=1(alg column)/0(alf row)) --> todelte columns

df.values -->covert df to numpy ARRAY
np.array(df)




-----------------NUMPY-------------------------------
2d vector space

x = np.array(x)

x = [
      [1,2,3],
      [4,5,6]
    ]

x[0]  --> [1,2,3] or x[:][0]
x[0][0] --.[1]

x[:] or x

x[::,2]  --> This outputs the 3rd column   x[row,column]

x[::,1:4]  -->This output the column values from 1 to 3rd column for all rows

x.reshape(-1,2) --> -1 adjusts the shape according to other dimensions
x.shape




----------------MATPLOTLIB----------------------------------------------





----------------------  TENSORFLOW ----------------------------------------

# Create a variable.
w = tf.Variable(<initial-value>, name=<optional-name>)

tf.placeholder(dtype, shape=None, name=None)

tf.truncated_normal(shape, mean=0.0, stddev=1.0, dtype=tf.float32, seed=None, name=None)

tf.zeros(shape, dtype=tf.float32, name=None)

tf.nn.relu(features, name=None)
features: A Tensor. Must be one of the following types: float32, float64, int32, int64, uint8, int16, int8, uint16, half.

tf.matmul(mat1,mat2)

tf.nn.sparse_softmax_cross_entropy_with_logits(logits, labels, name=None)
logits must have the shape [batch_size, num_classes] and dtype float32 or float64.
labels must have the shape [batch_size] and dtype int32 or int64.

tf.scalar_summary(tags, values, collections=None, name=None)
tags: A string Tensor. Tags for the summaries.
values: A real numeric Tensor. Values for the summaries.
collections: Optional list of graph collections keys. The new summary op is added to these collections. Defaults to [GraphKeys.SUMMARIES].
name: A name for the operation (optional).

tf.train.GradientDescentOptimizer(learning_rate, use_locking=False, name='GradientDescent')
learning_rate: A Tensor or a floating point value. The learning rate to use.
use_locking: If True use locks for update operations.
name: Optional name prefix for the operations created when applying gradients. Defaults to "GradientDescent".


sess = tf.Session(graph = )
  computed_x = sess.run(x)
              or
  computed_x = x.eval()

sess = tf.InteractiveSession()

g = tf.get_default_graph() ------>defaut graph in tensorflow all operation added to it

g1 = tf.Graph()

g1.get_operations().name

with g1.as_default():
    c = tf.constant(30.0)

w = g.get_tensor_by_name('layer1/W:0')
with tf.name_scope(name):
    a = tf.convert_to_tensor(a, name="a") -->name/a:0
    b = tf.convert_to_tensor(b, name="b")
--------convolutions--------------------------------

tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None, data_format=None, name=None)

input = [batch, in_height, in_width, in_channels]
filter = [filter_height, filter_width, in_channels, out_channels]
strides  = [1,1,1,1] or [1,2,2,1]
[input,stride_height,stride_width,channel]
padding = 'VALID'-> size will change
          'SAME' -> 0 padded to retain n the original shape