Maile Sachin Kumar

18BEI0006
L29+30

EXPERIMENT -1
DATA PREPROCESSING
Title of the Experiment : Data Pre-Processing

Aim : To implement the following data pre-processing techniques that can be applied
on data set to produce data for processing algorithms –
1. Scaling
2. Normalization
a. L1 Normalization
b. L2 Normalization
3. Binarization
4. Standardization
5. Data Labelling

Platform / Software used : Jupyter Notebook Online

Task 1:
Rescale to a range 0-10 and Display first 10 rows of the fi
le “pollution.csv”
Task 2 :
Normalize using L1 & L2 norms and Display first 10 rows of
the file “pollution.csv”
(i) Using L1 norms :

(ii) Using L2 norms

Task 3:
Rescale and Normalize using L1 & L2 and Display the result.
[Kindly download .csv dataset on your own from internet sou
rce]
(i) Rescale :

(ii) L1 Norm :
 (iii) L2 Norm :

Task 4 :
Binarize the data using a threshold of water usage value 475 and Display first 10
rows of the file
“yearly-water-usage.csv”
Task -5 :
Standardize the data in “yearly-water-usage.csv” file and plot the
water usage values (both normalized and original).
1. CODE AND OUTPUT AFTER STANDARDIZATION USING ORDINARY VALUES :
 GRAPH AFTER STANDARDIZATION WITH THE ORDINARY VALUES :

2. CODE AND OUTPUT OF NORMALIZATION:

CODE AND OUTPUT AFTER STANDARDIZATION USING NORMALIZED VALUES :

GRAPH AFTER STANDARDIZATION USING NORMALIZED VALUES :

Task 6 :
Repeat the above two tasks by using “pollution.csv”
1. BINERIZATION OF DATA :

2. CODE AND OUTPUT AFTER STANDARDIZATION USING ORDINARY VALUES :

Maile Sachin Kumar 18BEI0006

3. CODE AND OUTPUT AFTER STANDARDIZATION USING NORMALIZED

VALUES :