# Week $9-$ Exercise

### Notebook created by Jonathan Penava

For this exercise you are going to try to try and predict the carat range of a diamond based on some of its properties. Start by importing the standard libraries and reading in the 'diamonds.csv$\text{'}$ file.

Drop the 'Color', 'Clarity', and 'Unnamed: $0\text{'}$ Columns.

You will need to change the 'cut' values into int values.

Referencing what we did in week $5,$ create a new column for carat ranges with the values $\text{'}0-1\text{'},\text{'}1-2\text{'},\text{'}2-3\text{'},\text{'}3-4\text{'},\text{'}4-5\text{'},\text{'}5-6\text{'}$

Drop the column 'carat' after you have created the carat ranges column

Using K$-$Nearest Neighbours classify your carat ranges based on the other values in your dataframe. Print out an accuracy score and a classification report for your algorithm.

Create a diagram of error rates for different values of K$.$ What value of K will give you a more accurate result?

import pandas as pd

from sklearn.model$\_$selection import train$\_$test$\_$split

from sklearn.neighbors import KNeighborsClassifier

from sklearn.metrics import accuracy$\_$score, classification$\_$report

import matplotlib.pyplot as plt

diamonds$\_$df $=$ pd$.$read$\_$csv$(\text{'}$diamonds$.$csv$\text{'})$

diamonds$\_$df$.$drop$($columns$=[\text{'}$color$\text{'},$ 'clarity', 'Unnamed: $0\text{'}],$ inplace$=$True$)$

cut$\_$mapping $=\{\text{'}$Fair$\text{'}$: $1,$ 'Good': $2,$ 'Very Good': $3,$ 'Premium': $4,$ 'Ideal': $5\}$

diamonds$\_$df$[\text{'}$cut$\text{'}]=$ diamonds$\_$df$[\text{'}$cut$\text{'}].$map$($cut$\_$mapping$)$

carat$\_$bins $=[0,1,2,3,4,5,6]$

carat$\_$labels $=[\text{'}0-1\text{'},\text{'}1-2\text{'},\text{'}2-3\text{'},\text{'}3-4\text{'},\text{'}4-5\text{'},\text{'}5-6\text{'}]$

diamonds$\_$df$[\text{'}$carat$\_$range'$]=$ pd$.$cut$($diamonds$\_$df$[\text{'}$carat$\text{'}],$ bins$=$carat$\_$bins, labels$=$carat$\_$labels$)$

diamonds$\_$df$.$drop$($columns$=[\text{'}$carat$\text{'}],$ inplace$=$True$)$

X $=$ diamonds$\_$df$.$drop$($columns$=[\text{'}$carat$\_$range'$])$

y $=$ diamonds$\_$df$[\text{'}$carat$\_$range'$]$

X$\_$train, X$\_$test, y$\_$train, y$\_$test $=$ train$\_$test$\_$split$($X$,$ y$,$ test$\_$size$=0\mathrm{.}2,$ random$\_$state$=42)$

knn $=$ KNeighborsClassifier$($n$\_$neighbors$=5)$ # You can change the value of K

knn$.$fit$($X$\_$train, y$\_$train$)$

y$\_$pred $=$ knn$.$predict$($X$\_$test$)$

accuracy $=$ accuracy$\_$score$($y$\_$test, y$\_$pred$)$

print$($f$"$Accuracy: $\{$accuracy:$\mathrm{.}2$f$\}")$

print$($classification$\_$report$($y$\_$test, y$\_$pred$))$

error$\_$rates $=[]$

k$\_$values $=$ range$(1,21)$

for k in k$\_$values:

knn $=$ KNeighborsClassifier$($n$\_$neighbors$=$k$)$

knn$.$fit$($X$\_$train, y$\_$train$)$

y$\_$pred $=$ knn$.$predict$($X$\_$test$)$

error$\_$rate $=1-$ accuracy$\_$score$($y$\_$test, y$\_$pred$)$

error$\_$rates.append$($error$\_$rate$)$

plt$.$plot$($k$\_$values, error$\_$rates, marker$=\text{'}$o$\text{'})$

plt$.$xlabel$(\text{'}$K Value'$)$

plt$.$ylabel$(\text{'}$Error Rate'$)$

plt$.$title$(\text{'}$Error Rates for Different K Values'$)$

plt$.$show$()$ According to the above week $9$ exercise do the following

Week $10-$ Exercise

Notebook created by Jonathan Penava

B

For this exercise you are going to modify your week $9$ exercise to be a web service.

Have a sample URL that will return an example of JSON input.

Have an evaluate URL that will return the carat range based on the input.

Create a separate notebook $($Notes Part $3)$ that will test out your URL's