Contenido del Curso

Preprocessing Data

1. Data Exploration

Data Extracting Explore the Dataset Data Types

2. Data Cleaning

Missing and Wrong Data Work with NaNs Replace Numerical Missing Data with Values Dealing with Missing Ages Replace Categorical Missing Data with Values Replace Missing Values with Interpolation Removing Duplicates Removing Outliers Visualizing Outliers

3. Data Validation

Types Conversion Data Binning

4. Normalization & Standardization

Data Normalization Data Standardization

5. Data Encoding

Label Encoding One Hot Encoding

One Hot Encoding

One-hot encoding is one more preprocessing approach that is used before the training process. You already know about the LabelEncoding that transforms like this:

Embarked		Label
Q		3
S		2
S	->	2
S		2
C		1

To provide the model to process only 0 and 1 values, one hot encoder transforms to the matrix:

Embarked		C	S	Q
Q		0	0	1
S		0	1	0
S	->	0	1	0
S		0	1	0
C		1	0	0

1 means that the value of embark_town matches the following column name (for example, Queenston matches Q), and 0 - it doesn't match. Instead of saving n values in range 0...n-1, we create n columns filled with 0 and 1.

One hot encoding is quite useful in case if the cell contains multiple values. For example, your dataset contains sentences and a list of eemotions with which the sentence is labeled. It is not c convenient format to work with, so we transform it:

emotion		anger	joy	love	neutral	sad
sad, neutral		0	0	0	1	1
love		0	0	1	0	0
love, joy	->	0	1	1	0	0
anger		1	0	0	0	0
neutral		0	0	0	1	0

We will use OneHotEncoder to create new features for the categorical columns of our dataset.

OneHotEncoder cannot process NaNs, so you have to preprocess them first.

The common syntax is next:


              12345678910
            
from sklearn.preprocessing import OneHotEncoder

# data is loaded already
# num_cols and cat_cols are created already

encoder = OneHotEncoder()
new_data = pd.DataFrame(encoder.fit_transform(data[cat_cols]).toarray())

# join new features to the dataset, but remove categorical features
data = data[num_cols].join(new_data)

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Cambia al escritorio para practicar en el mundo realContinúe desde donde se encuentra utilizando una de las siguientes opciones

¿Todo estuvo claro?

Sección 5. Capítulo 2

One Hot Encoding

One-hot encoding is one more preprocessing approach that is used before the training process. You already know about the LabelEncoding that transforms like this:

Embarked		Label
Q		3
S		2
S	->	2
S		2
C		1

To provide the model to process only 0 and 1 values, one hot encoder transforms to the matrix:

Embarked		C	S	Q
Q		0	0	1
S		0	1	0
S	->	0	1	0
S		0	1	0
C		1	0	0

emotion		anger	joy	love	neutral	sad
sad, neutral		0	0	0	1	1
love		0	0	1	0	0
love, joy	->	0	1	1	0	0
anger		1	0	0	0	0
neutral		0	0	0	1	0

We will use OneHotEncoder to create new features for the categorical columns of our dataset.

OneHotEncoder cannot process NaNs, so you have to preprocess them first.

The common syntax is next:


              12345678910
            
from sklearn.preprocessing import OneHotEncoder

# data is loaded already
# num_cols and cat_cols are created already

encoder = OneHotEncoder()
new_data = pd.DataFrame(encoder.fit_transform(data[cat_cols]).toarray())

# join new features to the dataset, but remove categorical features
data = data[num_cols].join(new_data)

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Cambia al escritorio para practicar en el mundo realContinúe desde donde se encuentra utilizando una de las siguientes opciones

¿Todo estuvo claro?

Sección 5. Capítulo 2

One Hot Encoding

One-hot encoding is one more preprocessing approach that is used before the training process. You already know about the LabelEncoding that transforms like this:

Embarked		Label
Q		3
S		2
S	->	2
S		2
C		1

To provide the model to process only 0 and 1 values, one hot encoder transforms to the matrix:

Embarked		C	S	Q
Q		0	0	1
S		0	1	0
S	->	0	1	0
S		0	1	0
C		1	0	0

emotion		anger	joy	love	neutral	sad
sad, neutral		0	0	0	1	1
love		0	0	1	0	0
love, joy	->	0	1	1	0	0
anger		1	0	0	0	0
neutral		0	0	0	1	0

We will use OneHotEncoder to create new features for the categorical columns of our dataset.

OneHotEncoder cannot process NaNs, so you have to preprocess them first.

The common syntax is next:


              12345678910
            
from sklearn.preprocessing import OneHotEncoder

# data is loaded already
# num_cols and cat_cols are created already

encoder = OneHotEncoder()
new_data = pd.DataFrame(encoder.fit_transform(data[cat_cols]).toarray())

# join new features to the dataset, but remove categorical features
data = data[num_cols].join(new_data)

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Cambia al escritorio para practicar en el mundo realContinúe desde donde se encuentra utilizando una de las siguientes opciones

¿Todo estuvo claro?

One-hot encoding is one more preprocessing approach that is used before the training process. You already know about the LabelEncoding that transforms like this:

Embarked		Label
Q		3
S		2
S	->	2
S		2
C		1

To provide the model to process only 0 and 1 values, one hot encoder transforms to the matrix:

Embarked		C	S	Q
Q		0	0	1
S		0	1	0
S	->	0	1	0
S		0	1	0
C		1	0	0

emotion		anger	joy	love	neutral	sad
sad, neutral		0	0	0	1	1
love		0	0	1	0	0
love, joy	->	0	1	1	0	0
anger		1	0	0	0	0
neutral		0	0	0	1	0

We will use OneHotEncoder to create new features for the categorical columns of our dataset.

OneHotEncoder cannot process NaNs, so you have to preprocess them first.

The common syntax is next:


              12345678910
            
from sklearn.preprocessing import OneHotEncoder

# data is loaded already
# num_cols and cat_cols are created already

encoder = OneHotEncoder()
new_data = pd.DataFrame(encoder.fit_transform(data[cat_cols]).toarray())

# join new features to the dataset, but remove categorical features
data = data[num_cols].join(new_data)

Tarea

Apply the One Hot Encoding to the dataset.

Load the dataset.
Process the NaNs: drop it for the Embarked, replace with mean value for Age.
Transform the Cabin data as in the previous chapter (apply the Label Encoding).
Create the variable cat_cols to store such a categorical features: Sex, Cabin, and Embarked.
Create OneHotEncoder and store the transformed data to the new_data.
Remove the cat_cols from the dataframe, but add the new_data.
Check the sample.

Cambia al escritorio para practicar en el mundo realContinúe desde donde se encuentra utilizando una de las siguientes opciones

Sección 5. Capítulo 2

Cambia al escritorio para practicar en el mundo realContinúe desde donde se encuentra utilizando una de las siguientes opciones