Assignment[Optional]

Instructor: Applied AI Course Duration: 30 mins

Please download the data from https://www.kaggle.com/mlg-ulb/creditcardfraud

1. Please complete the videos of EDA and Linear algebra

2. Info about data: it is a CSV file, contains 31 features, the last feature is used to classify the transaction whether it is a fraud or not

Task 1: please do proper analysis of the whole data, plot all relevant plots, note down all observations.

Task 2: Let's define a matric similarity(i,j) = cosine^-1(dot product (vi, vj) / (length(vi) * length(vj)) ) Take out any sample from the data set which contains no less than 100 transactions, for every transaction in the sample find out top 10 transactions in the dataset which have the lowest similarity(i,j).

Note: the sample should contain the similar distribution of classes as the original dataset

ex: Sample S = {1, 34, 989, 1000} given transaction id = 1 , class = 1

-------------------

similar transactions

-------------------

class = 1 # similarity = 0.0023 # transaction id = 0

class = 1 # similarity = 0.02312 # transaction id = 89

class = 1 # similarity = 0.02498 # transaction id = 2034 ...

--------------------------------------------------------------------------

given transaction id = 34, class = 0

-------------------

similar transactions

-------------------

class = 0 # similarity = 1.005 # transaction id = 523212

class = 1 # similarity = 1. 2673 # transaction id = 500001

class = 1 # similarity = 1. 3459 # transaction id = 24

Note: Upload both python notebook and pdf version of that notebook

0 Comment(s)

4.16
14 min
4.30
19 min