Classification

$30$ Points

Using dataset Red$\_$wine$\_$quality $.$csv answer following questions. Either use Python coding or Weka. Provide your work in form of code and its output, if using Python. Screenshots, parameters used $($with explanation$)$ for Weka.

$1.$ Is the dataset balanced? Which technique did you use to resolve the issue $($Evaluation Method, Sampling technique, Threshold moving$).-5$Points

$2.$ Does dataset have Scaling issue? Did you normalize your data? Show your normalization of dataset. $-5$ Points

Use following code for normalization$-$

from sklearn.preprocessing import MinMaxScaler

scaler $=$ MinMaxScaler$($feature$\_$range$=(0,1))$

normal$\_$df $=$ scaler.fit$\_$transform$($df$)$

normal$\_$df $=$ pd$.$DataFrame$($normal$\_$df$,$ columns $=$ df$.$columns$)$

$3.$ What is dependent variable and what is its data type? Is it in right format? If not convert it to nominal datatype. $-5$ Points

$4.$ Create baseline model using following code$--5$ Points

from sklearn.dummy import DummyClassifier

dummy$\_$classifier $=$ DummyClassifier$($strategy$=$'most$\_$frequent',random$\_$state$=0)$

dummy$\_$classifier.fit$($X$\_$train,y$\_$train$)$

acc$\_$baseline $=$ dummy$\_$classifier.score$($X$\_$test,y$\_$test$)$

print$("$Baseline Accuracy $=",$ acc$\_$baseline$)$

$5.$ Choose one Classifier $($Decision tree, Na$$ve Bayes, or SVM$)$ to train your model. What is its accuracy, CV Score and F$1$ Score. $-5$ Points

$6.$ Is your model better than baseline model? How do you evaluate your model performance? $5$ Points