SIMPLE PREDICTION MODEL
Mean
Suppose you have given sales data of a certain commodity of past years and someone asks you to predict the sale for this year. Without any knowledge of machine learning algorithm, you can say that it will be approximate average/mean sales of past few years.
Mean is computed by dividing the sum of all values by the number of values.
It's a good start but it's also important to know the accuracy of our model. There are many methods which we can use to evaluate how good is our model. Most commonly used method is mean square error (MSE).
Mean Square Error
The simple way to calculate mean square error is calculating the difference between actual values and the predicted value. However, we can't add it simply because the positive and the negative error may cancel each other so square these error before adding. And divide the sum by the number of data point to calculate the mean error.
It's always important to calculate the error in our model so that we efficiently use resources. Let's implement this model in python.
Installation
sudo apt-get install python3-pip
sudo pip3 install numpy
sudo -H pip3 install pandas
Source Code
import pandas as pd
import numpy as np
file = path_of_your_csv_data_file
#read csv
df = pd.read_csv(file)
#convert data frame to numpy array
df = df.values
# select amount column from df, here it is in 3rd colum i.e 2
amount = df[:, 2]
# calcaulate average amount of sale in past year
mean_ = amount.mean(axis=0)
print ('Mean ', mean_)
#calculate mse
mse = ((mean_ - amount)**2).mean(axis=0)
print ('mse', mse)
Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above
Comments
Post a Comment