Kaggle Dataset Analysis : Is your Avocado organic or not?

Hey readers! Today, allow me to present you yet another dataset analysis of a rather gluttony topic, namely Avocado price analysis.

This Data set represents the historical data on avocado prices and sales volume in multiple US markets.

Our prime objectives will be to visualize the dataset, pre-process it and ultimately test multiple sklearn classifiers to checkout which one gives us the best confidence and accuracy for our Avocado's Organic assurance!

Note : I'd like to extend the kernel contribution to Shivam Negi. All this code belongs to him.

Data Visualization

This script must procure the following jointplot 

sns.jointplot(x='Large Bags',y='Small Bags',data=avocado)

view raw bagsjp.py hosted with ❤ by GitHub

While a similar joint plot can be drawn for conluding the linearly exponent relations between extra large bags and the small ones.

sns.jointplot(x='Small Bags',y='XLarge Bags',data=avocado)

view raw bagsjp2.py hosted with ❤ by GitHub

Pre Processing

The following script has been used for pre processing the input data.

	from sklearn.feature_extraction import FeatureHasher
	from sklearn.model_selection import train_test_split
	from sklearn.neighbors import KNeighborsClassifier
	from sklearn.metrics import classification_report,confusion_matrix
	from sklearn.ensemble import RandomForestClassifier
	from sklearn.svm import SVC
	from sklearn.grid_search import GridSearchCV
	avocado = avocado.drop(['Unnamed: 0','Date'],axis=1)
	# Using Feature Hashing
	fh = FeatureHasher(n_features=5,input_type='string')
	hashed_features = fh.fit_transform(avocado['region']).toarray()
	avocado = pd.concat([avocado,pd.DataFrame(hashed_features)],axis=1)
	avocado = avocado.drop('region',axis=1)
	#Simple Train test splits
	X = avocado.drop('type',axis=1)
	y = avocado['type']
	X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)

view raw preprocessing.py hosted with ❤ by GitHub

Model Definition and Comparisons

We will be looking mostly at three different models, namely random forest classifiers, KNN and our good old SVM.

Obviously, the Grid search optimization has been used for best results.

Random Forest Classifier

	rfc = RandomForestClassifier(n_estimators=100)
	rfc.fit(X_train,y_train)
	pred1 = rfc.predict(X_test)
	#Classification reports
	print(classification_report(y_test,pred1))
	#Confusion Matrix
	print(confusion_matrix(y_test,pred1))

view raw randomforest.py hosted with ❤ by GitHub

You can see that our model is giving (false positive) + (false negatives) = 3. Total 3 inaccurate predictions which is pretty awesome.

KNN Classifier

	knn = KNeighborsClassifier(n_neighbors=5)
	knn.fit(X_train,y_train)
	pred2 = knn.predict(X_test)
	#Classification Report
	print(classification_report(y_test,pred2))
	#Confusion Matrix
	print(confusion_matrix(y_test,pred2))

view raw knn.py hosted with ❤ by GitHub

So, the performance drops by a total 0f 0.02, not bad at all!
For The SVM Part, I'd recommend that you visit the kernel itself (Link is at the top)

Conclusion

In the end, I would just like to say that we have tried Random Forest Classifier, KNN and SVM but most efficient among them is Random Forest Classifier giving 100% accuracy, KNN is also not too bad as it has 98% accuracy but Support Vector Classifier(SVC) is not that much efficient in predicting the values having only 54% accuracy for conventional and 100% accuracy for organic type. Recommended Model for prediction of TYPE(coventional and organic categories) columns is Random Forest Classifier.