Great demonstration. My concepts are cleared now regarding how to fit linear regression on Such kind of datasets where there are multiple attributes like fuel kms used engine power transmission.
How can we change the colour of Predicted Prices in the scatter? This will help us understand better the difference between Actual Prices which will be visualized in blue colour and Predicted Prices which will be visualized in red colour.
Thanks for a newbie friendly video. I have 2 questions. 1. A newbie question. How to deploy this as a production model to predict a car price for a given set of inputs? 2. How to plot Actual Prices vs Predicted Prices in different colors?
@@Siddhardhan Hi, I also trying to different plot colors for Actual Prices vs Predict Prices. When I tried to add hue I am getting an error. How exactly can we implement this?
Data Collection and Preprocessing Feature of Engineering and Selection Model Selection and Evaluation Model training and testing Result and analysis Conclusion
Would it be beneficial to perform one-hot encoding on the categorical features that have more than 2 categories (e.g. 0, 1, 2...)? It was suggested in a different lecture, but i dont know what to do then if the number of categories go higher like 10 or more.
For this case, I do not think it would be beneficial as it will increase memory consumption. What I am concerned about is whether should we use a standard scaler here.
The Dataset Can't be accessed any more. The second time i am encountering this issue with the projects. Consider this when making your projects because i believe the projects are made to educate for a long time. Thank You
Thanks for making this impressive vdo tutorial alot, sir . I am now working on this project for my school assignment, but since I am a newbie to ML. I am wondering whether there is any possible way to make damaged car identification using machine learning, also it can predict, what status our cars are in ( good, new, .. ) and if good how many percentages. Looking forward to hearing back from you thanks
bro, while you used lasso at end, you got test data r2 score is more compare to train data r2 score. How we can infer on this? is this overfitting or under fitting?
to test the validity of model beyond r2 score if we specify attribute values that are closer to one of the training data record, in the final output window, should the model need to predict the price nearer to selling price for that specific record in training data? Expecting your answers. Thanks
In these 2 Projects, I have used Logistic Regression: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-fiz1ORTBGpY.html ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nacLBdyG6jE.html
Thank you for informative video. Why didn't you use dummy variables instead of manually doing it yourself? Is it just a preference or there are advantages to it?
@@Siddhardhan oh okay, awesome. May you please assist me with writing a function that; - Takes in a dataframe (df), - Standardize the features using StandardScaler() - Split data into 25% training, random_state =25% - Fills any NAN values with zero. I’m able to do it step by step but not as one function
hi! you can try and.lmplot() from seaborn to plot linear model with your data. Refer seaborn documentation on how to do this. we can find the difference of the actual price & predicted price with metrics like mean absolute error, etc.
@@Siddhardhan ok thanks,did both of it,but how this model worked behind ,which feature(column) have been given how much importance,for predicting selling price,please suggest best book for maths for machine learning,i am new to this field
33:55 the Rsquared. When do we use this? Is it because using linear model? is it only for linear model? is there any spesific reason sir on when exactly we use this metrics?
When we print(car_dataset.Fuel_Type.value_counts()), it shows Name: Fuel_Type, dtype: int64. Does this int64 imply no. of counts? (As Fuel_Type column contains, string type data, I got confused)
I have confusion on using same dataset for training and prediction. I think you should use test dataset for prediction instead of training dataset i.e. X_train. Please clarify my doubt.
hi! the reason for finding the accuracy on x_train and x_test is to determine whether the model has overtrained. that's the reason why I used x_train as well.
Superb work! Splitting the data into train and test---Even after spelling correctly, I get an error "the column not found in axis'' Can I replace data.drop with data.iloc for all cases.
Hi! I'll be posting a video on feature extraction of text data this Wednesday. You can watch this project video. I have explained how to convert text to feature vectors. (Fake News Prediction) : ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nacLBdyG6jE.html
how to find 12. Which is the most and least cost depreciated vehicle in data ? 13. Which brands of vehicles are less affected by cost depreciation ? 14. Can we get idea about newest vehicles i.e. after 2014 manufactured ? 15. Can we find out data of only two wheelers from this data? pls answer this fast pls
@@Siddhardhan is it enough If I just make a model in Jupiter notebook..as I don't enough time further steps like deployment as I want to include atleast one project in ML in my resume...is it enough sir
While writing the encoding of data using replace function, I'm getting an error "TypeError: Cannot compare types 'ndarray(dtype=int64)' and 'str' ", so I'm not able to change my values to 0,1 & 2.
The reason for this error is because of the number of times you execute the cell. The first time you execute the cell the transformation happens ('Y' -> 1, and 'N' -> 0). At this point if you try to rerun the same cell, the data type for physicalEvidence and contact is now an int rather than a string. So when the replace method runs, it is look for a 'Y' or 'N', but DOES NOT find any because all of the values have been replaced with 1 or 0. If this happens, you'll just need to reload the dataframe (df) again and rerun the cell one time. I hope this helps! Let me know your thoughts.