Stock market scenario: Case 1 : we have predicted it will crash ,but it did not (False Positive) . Focus on precision score. Case 2 : we have predicted it will not crash ,but it crashed (False Negative) . This will do more damage than case 1. Hence, we need to Keep FN low (high Recall) . Case 3: If client specifically wants a balance i.e we don't want too many FP or FN . Focus on F1 score.
Hello sir, in response to the question asked,here is my answer. so the problem statement is to predict whether the stock market will crash or not. here the FALSE POSITIVE would be/Type 1 error rate would be: our model predicts that the stock market crashes but in reality it does not crash(predicting something true when in reality it is false). here the FALSE NEGATIVE would be/Type 2 error rate would be: our model predicts that stock market wont crash but in reality it crashes(predicting something false when in reality it is true). so in context of the stakeholder or someone who is really into buying stocks,that person would try to minimise the False Negative rate as it brings much more harm to that person, as a lot of money is attached to the stock market,perhaps.
Really Great !!!!Are you going to release such beautiful videos with your suggestions for those who are going to join as a Team Lead or Data Science Managerial Role.
Hey any good research paper on stock market fraud detection...like circular trading nd spoofing...?.. I need to develop it as my use case...if u know plz mention some solved use case for reference
If the interviewer asks you, tell me about yourself, then start with the best project that you had done and have an end to end knowledge of it. because the next question they ask will from that project only and as you have an end to end knowledge of it, you can answer it easily.it also help you to build confidence.
If we talk about whether the stock market will crash or not, shouldn't we foucs on reducing false negatives as FN are cases where the predicted result was falsely labeled as negative and in this case, if the stock market crashes and the model predict that it doesn't crash, it will make no sense. Kindly correct me if you have other analogies
Stock market scenario:
Case 1 : we have predicted it will crash ,but it did not (False Positive) . Focus on precision score.
Case 2 : we have predicted it will not crash ,but it crashed (False Negative) .
This will do more damage than case 1.
Hence, we need to Keep FN low (high Recall) .
Case 3: If client specifically wants a balance i.e we don't want too many FP or FN . Focus on F1 score.
It's 💯 true!! I have also cleared my interview by following him❤️
Is feature selection techniques discussed in the channel
Congrats👏👏
Can u share some interview questions they asked for u? It heps me lot. I am going to attend
Hello sir, in response to the question asked,here is my answer.
so the problem statement is to predict whether the stock market will crash or not.
here the FALSE POSITIVE would be/Type 1 error rate would be: our model predicts that the stock market crashes but in reality it does not crash(predicting something true when in reality it is false).
here the FALSE NEGATIVE would be/Type 2 error rate would be: our model predicts that stock market wont crash but in reality it crashes(predicting something false when in reality it is true).
so in context of the stakeholder or someone who is really into buying stocks,that person would try to minimise the False Negative rate as it brings much more harm to that person, as a lot of money is attached to the stock market,perhaps.
Thank you so much for your valuable advice in this tough time. Much needed video
Eagerly Waiting for this one
Excited, waiting for the video
Really Great !!!!Are you going to release such beautiful videos with your suggestions for those who are going to join as a Team Lead or Data Science Managerial Role.
Hi Krish... Pls do one live streaming on unsupervised clustering algo end to end implementation.
What a motivation u have gave👏🙏🏿kudos
Great Videos, You are really doing a great job.
can u make a project on hand gestures recognition based software control using cnn ?
Thanks Krish
Hey any good research paper on stock market fraud detection...like circular trading nd spoofing...?.. I need to develop it as my use case...if u know plz mention some solved use case for reference
Pincode I used Probability Encoding Technique
Thank you so much sir 🥰
Can u talk data pipelines in story telling ...how data we are getting.
Nice info.
stock market : False Positive should reduce , i.e. a rising stock price, if predicted with a fall in price would do more harm
waiting...............
Is Feature selection techniques discussed in the channel Krish?
I'm not able to find it
@@PraveenKumar-pd9sx There is a playlist named Feature Enginnering where this topic is covered
Hi searched for that but could only see Encoding outlier detection and imbalance handling and not featire selection techniques
Can you tell me a good resource for that
@@PraveenKumar-pd9sx ruclips.net/video/EqLBAmtKMnQ/видео.html Here is Krish video !
Krish bro how could we introduce our data science portfolio in interview
If the interviewer asks you, tell me about yourself, then start with the best project that you had done and have an end to end knowledge of it. because the next question they ask will from that project only and as you have an end to end knowledge of it, you can answer it easily.it also help you to build confidence.
Hello Sir How are you
Which will be the most demanding & most high paid job in coming 2-3years
Hi Krish, I am a software Quality Assurance analyst with 7 years of experience. Can I join ml project?
Stock market false positive
If we talk about whether the stock market will crash or not, shouldn't we foucs on reducing false negatives as FN are cases where the predicted result was falsely labeled as negative and in this case, if the stock market crashes and the model predict that it doesn't crash, it will make no sense. Kindly correct me if you have other analogies
@@kaustubhgupta12 please visit Kris video of performance metrics interview questions comments section, you get a better idea, may be l am wrong....