CODEWITHSUNDEEP
pythonmachine-learningscikit-learnlogistic-regressionprediction
Ravinder
Ravinder

Reputation: 11

sklearn Logistic Regression - Make prediction with custom input from outside testing data

I'm doing Logistic Regression with sklearn to predict some categories given some descriptions. Here's the code at the moment 

    X_Train, X_Test, y_train, y_test = train_test_split(df['description'], df['category'])

    count_vect = CountVectorizer()
    X_Train_counts = count_vect.fit_transform(X_Train)

    tfidf_transformer = TfidfTransformer()
    X_Train_tfidf = tfidf_transformer.fit_transform(X_Train_counts)

    # Fit the logistic regression model
    clf = LogisticRegression(random_state=0, class_weight='balanced', solver='lbfgs', max_iter=1000)
    clf.fit(X_Train_tfidf, y_train)

To also manually check my predictions I do this

    # Make predictions
    predictions = clf.predict(tfidf_transformer.transform(count_vect.transform(X_Test)))

    print(X_Test.iloc[7])
    print(predictions[7])

My question is how can I make a prediction of a category by manually giving a custom description from outside the testing data (e.g. I manually input it)

If that's possible, is there also a way to get the top n predictions categories for that custom text?

Upvotes: 1

Views: 528

Answers (1)

Andrey Lukyanenko
Andrey Lukyanenko

Reputation: 3851

You should be able to do something like this:

your_description = "some text"
vectorized = tfidf_transformer.transform(count_vect.transform([your_description]))
predictions = clf.predict(vectorized.reshape(-1, 1))

Upvotes: 1

Related Questions