Nullval

This repository contains the required package containing various mathematical
approaches using different numerical technique

Under construction! Not ready for use yet! Currently experimenting and planning!

Developed by Mukul namagiri

• This repository contains different kinds of methods for the treament of null values and outliers
  Using various kinds of numerical techniques for the ideal replacement of values in your dataframe

Accepted format

• This module takes xml, json, csv and excel and pandas dataframe as input
• automatically identifies the locations of null values and outliers
• ideal values for data imputations

Directory structure of the repository

nullvalue/
│
├── .gitignore
│
├── nullval/
│ ├── __init__.py
│ ├── cubic_spline_interpolation.py
│ ├── linear_interpolation.py
│ └── loader.py
| |__ polynomial_interpolation.py
| |__ splines_interpolation.py
| |__ trigonometric_interpolation.py
| |__ auto.py
│
├── tests/
│ ├── init.py
│ └── test_lagrange_interpolation.py
| |__ test_linear_interpolation.py
| |__ test_polynomial_interpolation.py
| |__ test_spline_interpolation.py
| |__ test_trigonometric_interpolation.py
│
├── api_reference.md
│
├── pyproject.toml
│
├── README.rst
│
└── README.md

requirements for the package

They are already added to the toml file but in case

pandas==1.3.3
numpy==1.21.4
tqdm
scikit-learn==0.24.2
seaborn==0.11.2
matplotlib==3.5.1
statsmodels==0.13.0
tensorflow==2.8.0
plotly==5.5.0

Installation

pip install nulval

Usage guide

loader loads and formats the data and auto fins the ideal solution

Step - 1

from nullval import loader

path = "<enter the default path according to the environment>"
# converts to dataframe
data = loader.auto(path)
# returns the index of the nulls and the outliers 
loader.nulls_and_outs(data)

Advantages and the Disadvantages of each of the method

Linear interpolation

Advantages

• Easy to implement and less computational requirements
• Quick to compute and effective for larger data sets with loads of missing values
• have more local control, less sensitive to outliers, works well with noisy data, handles discontinous data well

Disadvantages

not good for complex patterns, sharp corners, poor performance for smooth functions, requires higher order derivatives

Lagrange interpolation

• Straight forward, tries to give the best fit
• works for equidistant and the non equidistant points, no need to solve linear systems

Disadvantages

Runge's phenomenon for higher degree and the widely spaced points --> oscillations occur at edges of intervals leading to poor approximation higher computational costs and does not work for dynamic dataset, higher storage requirements

Splines interpolation

Advantages

• gives more local control by breaking down the domain into smaller fragments, more precise interpolation
• smoother interpolation and reduces oscillations, differentiable, piecewise continous

Disadvantages

More computataional effort, hard to choose appropriate boundaries, could lead to overfitting, takes significant resources, higher memory usage, beyond range interpolation

Polynomial interpolation

Advantages

• gives the exact fit, provides analytical expression for further theoretical analysis
• allows for flexibility in choosing the base polynomial

Disadvantages

same as those of lagrange

Trigonometric interpolation

Advantages

• Most natural fit for periodic data and capture harmonics well, gives high precision for smooth functions
• avoids runge phenomenon, fast computation with fft and basis function

Disadvantages

non periodic data issues, discontinous boundary effects, global nature