Pre-processing the data before applying a Machine Learning Algorithm

  • Rescale data
  • Standardize data
  • normalize data
  • Barnarize data

Rescale Data: 

  • Rescale all the attributes to have the same scale. Generally attributes are often rescaled into the range between 0 and 1 for better optimization.
  • We can rescale the data using scikit-learn using the MinMaxScaler class.









Standardize data:

  • Standardization is a useful technique to transform all attributes to a standard Gaussian distribution with Mean 0 and Standard deviation 1 for better optimization. 
  • We can standardize the rescaled data using scikit-learn with the StandardScaler class.










Normalize Data:

  • Normalizing refers to rescaling each observation (row) to have a length of  1(unit norm).
  • We can Normalize data with scikit-learn using the Normalizer class. This pre-processing is used for sparse datasets(attributes having lot of zero values)










Binarize Data:

  • We can transform the data using the binary threshold. All values above the threshold are marked as 1 and all equal to or below are marked as 0. 
  • Making crisp values when we have probabilities as feature values.
  • We use a scikit-learn binarizer class

Comments

Post a Comment

Popular posts from this blog

Image
Feature selection for Machine Learning The data features used to train the machine learning model have a great impact on the ultimate performance. Irrelevant or partially relevant feature can negatively influence the model. The various automatic feature selection techniques are: Univariate Selection Recursive feature elimination Principal Component Analysis Feature Importance   Benefits of feature selection techniques: Reduces overfitting  Improves accuracy  Reduces training time Univariate Selection: This selection can be used to select the features that have the strongest relationship with the output variable. The example below uses scikit-learn which provides SeleceKBest class that can be used combinely with chi-squared (chi2) statistical test for non-negative features to select 4 of the best features from the dataset(Pima Indian). Recursive Feature Elimination: REF works by recursively removing the attributes a...
Read more
Image
Knowing sample dataset for use in Machine Learning The Pima Indians Diabetic Dataset is used to illustrate the Machine Learning concepts. The dataset describes the medical records for Pima Indians and whether or not each patient will have an onset of diabetes within five years. # Columns description    Pregnancies(preg) Number of times pregnant Glucose(plas) Plasma glucose concentration a 2 hours in an oral glucose tolerance test BloodPressure(pres) Diastolic blood pressure (mm Hg) SkinThickness(skin) Triceps skin fold thickness (mm) Insulin(test) 2-Hour serum insulin (mu U/ml) BMI(mass) Body mass index (weight in kg/(height in m)^2) DiabetesPedigreeFunction(pedi) Diabetes pedigree function Age(age) Age (years) Outcome(class) Class variable (0 or 1) 268 of 768 are 1, the others are 0 ...
Read more
Image
How to load Machine Learning Data The most common format to load machine learning data is CSV files. The most common ways to load the data in Python: Load CSV files using Python Standard Library. Load CSV files with NumPy. Load CSV files with Pandas.
Read more