lagrange.py

from typing import List
import numpy as np
import pandas as pd
import math

"""
        lagrange is a mathematical interpolation methods that approximate 
        a polynomial of degree n given n-1 points 
"""



def lagrange(X,Y) -> List:
    """
        the lagrange method that takes n points as input and return a pandas dataframe
    """
    my_columns = ['x','y']
    dataframe = pd.DataFrame(columns=my_columns)
    
    for pt in np.arange(-1,2,0.001):
        condition = False
        for i in range(0,len(X)):
            check = False 
            for j in range(0,len(X)):
                if (i is not j and check==False): 
                    Lp = (pt-X[j])/(X[i]-X[j])
                    L = Lp 
                    check = True 
                elif (i is not j):
                    L = L* (pt-X[j])/(X[i]-X[j])
            if (condition ==False):   
                P = Y[i]* L  
                condition = True 
            else: 
                P = Y[i]* L + P
        dataframe = dataframe.append(
            pd.Series(
                [pt,P],
                index=my_columns
            ),
            ignore_index=True)
    return dataframe 

if __name__=="__main__":
    """
        creating the polynome that we want to approximate 
        using lagrange method
    """
    my_columns = ['x','y']
    dataframe = pd.DataFrame(columns=my_columns)
    for i in np.arange(-1,2,0.001):
        dataframe = dataframe.append(
        pd.Series(
            [i,pow(i,3)],
            index=my_columns
        ),
        ignore_index=True)
    """
        calculating the mean square error from the given polynome and the result 
    """
    res = lagrange(np.array([-1,1,2]),np.array([-1,1,8]))
    m = len(dataframe)
    diff = pow(dataframe['y']-res['y'],2).sum()
    RMSE = 1/m * (math.sqrt(diff))
    print(RMSE)