Whether We Should Reduce False Positive Or Negative In Confusion Matrix-Machine Learning Interviews

⭐ Kite is a free AI-powered coding assistant that will help you code faster and smarter. The Kite plugin integrates with all the top editors and IDEs to give you smart completions and documentation while you’re typing. I've been using Kite for a few months and I love it! www.kite.com/g...
Subscribe my vlogging channel
/ @krishnaikhindi
Please donate if you want to support the channel through GPay UPID,
Gpay: krishnaik06@okicici
Telegram link: t.me/joinchat/...
Please join as a member in my channel to get additional benefits like materials in Data Science, live streaming for Members and many more
/ @krishnaik06
Connect with me here:
Twitter: / krishnaik06
Facebook: / krishnaik06
instagram: / krishnaik06

[@shilpaprusty3319]

False positive should be low when we are trying to convict a guilty person .An innocent person should never get punished.

[@simplesheenu]

Minority Report movie is based on this

[@ritik84629]

Prison break is based on this thing. May 100 guilty left unpublished but never torture 1 non guilty one!

[@hrishikeshnamboothiri.v.n2195]

Must reduce FN: If vaccine has side effects and predicted no side effect, the vaccination may lead to another health problem. Also we must reduce FP : IF the vaccine has no side effects and predicted, it has side effect, people won't be ready for vaccination and it will lead to more worse situation. Nowadays Black fungus also spreading among covid patients. So reducing both FN and FP is important.

[@mahantesh7678]

Spam detection: Where we should have lesser false positives compared to false negatives. Because it is ok to detect spam as not spam as a user can determine it as spam but the algorithm shouldn't detect important non spam mail as spam, as mail will be moved to spam folder and user miss it out

[@senthilkumarradhakrishnan744]

Best i have seen for this scenario... 👏

[@navyaatn3277]

Thank you krish naik! This video means a lot for me in a current real life scenario! I am really grateful. Thank you

[@karthikvegeta5981]

the place where you ll be concentrating more on false positive is when you need to sell a thing...lets take an example that I saw in sudhanshu n krish's interview ... he asked a question to concentrate on which FP or FN when a seller need to sell a washing machine whether it works or not ..obviously this is a use case of FP because when the situation come to selling the seller needs to sell even it is not working in this case having higher score on precision would definety helps you to yield more..but still everything depends on domain usecase...

[@RandomGuy-hi2jm]

Yeh that will definitely clear the doubts 👍

[@boejiden7093]

My friend got this question when interviewing at oracle. I couldnt answer him at the time. Now I understand. Thanks

[@PrashantKumar-mj8nb]

I think we need to reduce the false positive for the usecase(biometric system): Whether the particular person has access to the restricted area or not.

[@rangerover5635]

In the person having disease eg:- Cancer, the the False Positive is also very important as the chemotherapy on a healthy person could actually harm the person.

[@abde1581]

In vaccination case both FP and FN are important. 1). If vaccine really has an side effects but our model predict it hasn't, then it will leads to disaster. So, for that we need to reduce FP. 2). On other hand if vaccine doesn't have any side effects but model shows that it has. Then this results might be put a fear in people's mind and they will not go for vaccination. In this case FN become more important. So, good model that predict whether a vaccination has any side effects or not should have reduce both FN and FP results.

[@lastminuteprogrammer8511]

In the stock market example, one more reason to reduce False Positive is that: Imagine if a popular prediction website wrongly predicts market is going to crash tomorrow, then due to panic, many people will start selling stocks, which in turn will cause market to eventually go down (due to panic selling) even if originally market was not about to go down. This is why stock markets are hard to predict, as even our predictions can influence the market behavior in future.

[@vinodbiradar5266]

Exactly, market sentiments play a crucial role in this problem

[@janmoy4670]

So by the prediction, you can cause a self-fulling prophecy

[@bhargavikoti4208]

Sahiii👍👍..

[@Nektor_app]

Spam email detection false positives can make user loose/delay important data whereas false negative is still acceptable

[@nagnathsatav9978]

Use case is spam filter, FP should reduce because e-mail is not spam but predicted as spam.

[@vishvanathmetkari3444]

perfect

[@rajkir2852]

Well explained

[@shivalikrishna6565]

FP can be considered in examples below: 1. Prediction of demand of product in market. Postive Prediction- Reach demand of K units. FP- Company will manufacture far more than demand and ultimately results in loss of resources, money, time, man power. FN- If prediction is demand will not reach K units, And it turns out false, Company has option to quickly accelerate it's production. 2. Stock price prediction. Positive Prediction- Stock price of the firm will have a steep rise in future. Person will invest more money onto that and prediction turns out to be false means loss of money invested. FN- Prediction say Stock price will not rise steep. Person will not invest money on that. Atleast one cannot loose their money in FN, if not benefitted from that False Negative prediction.

[@isi6402]

In the first example you mentioned about demand and supply there is a scenario of FN in which Actually there is high demand but model Predicts less so Less production and less revenue as company not meet Required Value Of demands.

[@tauqeeralitapya6447]

One of the most helpful videos, Thanks Sir!

[@rajkir2852]

When doctor predicts i am covid negative but i was actually positive : FN. Daammm.... its risky for me and for my family

[@kuldeep27396]

@Krish Please try to make videos with white board. Dark mode affects eyes. There is a research on that, we should use less dark mode. White board will be helpful for all.

[@aliasfake]

Betting on a team for the win: here FN should be lower as if team is not win and it's predicting that team will win than we may loose our money. In vice versa situation atleast we are not losing the money.

[@balapranav5364]

Hi sir please make video on Boosting - CATBOOST Technique Unsupervised learning- t-sne and umap Please it will be more helpful for us

[@declanraj]

FP should be reduced more than FN when predicting if the stock market will increase. If FN (predict not increase but actually increase) at least I don’t lose money. But if FP (predict increase but actually decrease) I will lose the money.

[@rohitpanjabi4762]

i was just studying about ROC curve

[@davraheal6606]

Nice intro :)

[@lorizhuka6938]

Great

[@thepresistence5935]

intro love !

[@rakshitsinha4392]

A case where a model predicts whether I'll be able to clear the interview of a company or not. If it makes a False Positive case then the person will believe that they will get selected and will start relaxing/enjoying/partying and stop applying to other companies.

[@anuragdhadse1426]

Classifying a good Candidate for a job. In this example, Detecting a False Positive occurs when model classifies a not so good candidate to be absolutely perfect for the job position. This has very high consequences for the recruitor as it increase the time for the recruitment process to complete.

[@atharva-e5g]

FP shoould be low when we're recommending ads to specific person might be he is not interested in our products or FP should affect in the niche where demand is low but our model is showing that high demand and we are investing our money in that niche

[@nagnathsatav9978]

Actually positive but showing negative

[@aravinthmegnath3569]

Imagine a job consultancy has a ml applied. For a particular candidate the model predicts that he will have a job but in actuality he won't get it. In this case person will high probability of getting depressed or has to wait for a long time thinking that he will get a job. So we have to reduce the False Positive , so that our model will predict that the person will not get a job

[@lorizhuka6938]

In vaccinatin side effect problem we should try to reduce the false negative since if we predict that the vaccine will not have side affent and in reality it will have it may cause bad healthy consequences.

[@yesmom8594]

Reduce FP: When products have harmful content and hence cannot be consumed but predicting to be consumed will cause severe health issues

[@arunnandam4636]

In the case of vaccination, both the FP and FN should be reduced as In FN: Actually, there were No side effects, but our prediction shows there are side effects In FP: There are sideeffects and model predicting there are no sideeffects.

[@koustav6884]

FP should also be reduced in situation where climatic situation where acutally it is not going to rain our model is predicting that it is going to rain so we should reduce this type 1 error.

[@poojasethi5108]

Very nice video 👍👍

[@Nudaykumar]

Domain : Mining. Actual: Specific area has NO mines Prediction: Mine in such area. FP should decrease than FN , because of it lot of pollution gets created, waste resources utilization. Where as FN, no harm to environment if model predicts as no mines but mines exists.

[@mananpatel1700]

False-positive should be less while deciding the market value of any sports players/ weather there is going to be an extreme weather condition in coming days

[@yashrathi9793]

we can use youdens test to find the cutoff point to optimize false positive and false negetive

[@shaonsikder556]

When we should stop neural networks training?

[@naveenkumarjadi2915]

FP should be reduced in a person commited to crime or not

[@prasanthchowdaryvattikonda4829]

FP should be reduced: when a person is not vaccinated and model predicted as vaccinated. It is okay when the model is predicted as not vaccinated when a person already got vaccinated. Govt go and check-in records or green pass validation

[@rahuldevnath14792]

Krish i understood on when to reduce FP and FN, can you also help me with a live practical demo..how do we reduce that in logistic regression hands on...please!!!

[@chakhil8000]

may be credit card default prediction comes under scenario that we need to reduce false positive

[@aitoquantum]

In bating app whether a team will win or not. If it is predicted to win I will put my money but if it is FP I loose my money, on the other hand if it is is predicted to loose I dont put my money even if it wins my money is saved!

[@thulasikumar770]

Krish, make video of boosting family- light GBM, xgboost, adaboost and when to choose which one

[@nagnathsatav9978]

1.FN should be low

[@ratishr6003]

FP or Type 1 error is more bad use case : Predicting an Email as Spam or Ham

[@bigchallenge7987]

False-positive is important in case of covid-19 because if a person is corona negative and report comes corona positive then he may admitted to hospital and get corona as this virus spread socially by having contact who has covid.

[@sajaldhingra9321]

FP should be low incase the weather predicts of a gloomy weather

[@muhilarasan7212]

need to reduce FN and FP too.

[@krrsh]

May be an irony but whether India wins a final match or not can be considered for reducing False Positive, cos Fans won't be upset if it doesn't win...

[@trustmelabisfun3304]

Can u turn on the subtitle for this video plz? Thank you

[@shajahanpj4233]

Can a person survive while surfing on big waves? FP is Important

[@fariqjamil5484]

Examining the paper whether a person has cheated or not

[@ajithanbharathi6272]

Detecting the mail as spam or not spam

[@ilanaizelman3993]

🎉🎉🎉

[@nirajsathe9220]

FN should be low while we r unlocking the lock using face detection...

[@tavleenbajwa9773]

If we take an example of reaction chain going inside a fermenter which flows in the direction A to B in time t and then B + Z (Z will added after the formation of B), but if our model predicts a False positive that A to B will be formed. But in reality A forms some other element name D and we add Z after time t, then it can react with D instead of B and lead to a failed experiment and result in noxious chemical production, thus wasting time and resources. But we we have greater false negative, then we can definitely stop the experiment at the initial stage after time t to check if our experiment is going in the right direction or not.

[@abhimanyuyadav1387]

I am bit confused sir. if I am not wrong Type 1 error should be FN and Type 2 error should be FP....

[@arjj5232]

IDS or Security server alert that security breach occurred where it is not. Could it be taken as example of False positive?

[@bhawna1997]

Reducing False Positives - Accepting a claim for insurance, it will be incur loss to insurance company

[@ajaykushwaha4233]

Overall nichor of this lecture is FP and FN has to be low.

[@mdabuzar9300]

NO. It depends on the Use cases.. Blindly we can not go for minimizing FP and FN. It may cause feature loss, etc., and always there is a trade-off , so we have to keep in mind.

[@vageshyaggati8214]

FN+FP= F1 score

[@abhishekprakash9803]

how to reduced false positive or false negative, answer please

[@preetisri7297]

For FP whether an email is spam or not

[@1234560pratik]

you never ever use those words pretty much easy and pretty much simple guys 😂 because i watched many times to understand the whole concept..Really its CONFUSION Matrix...

[@YogeshKumar-tj3te]

how you can reduce false negative ?

[@SachinModi9]

critical alarm system in factory should have lesser false positive

[@arindamghosh3787]

In the vaccination case both FN and FP should be reduce . If FN is high ,then people are more likely to be suffer by the side effect of the vaccine . And if FP is high then the model will be predicting that vaccination has a side effect when it is not .Then in that scenario people will not take the vaccination and they are more likely to have the disease .So ,both our FN and FP should be low in this model

[@vishalbhapkar2359]

Vaccination problem, False negatives FN should be lowest

[@ravanshyam7653]

sir i think fp and fn are reversed there .i mean it swaps there orginal postion

[@arunnandam4636]

It is correct only, compare with definitions

[@kuldeep27396]

For vaccination FN should be low

[@rit_s__]

Is this course is good for placement

[@rit_s__]

If we compare krish naik KZbin free video and that course video , is their any content variation ??

[@rit_s__]

@@kuldeep27396 in compare with krish naik KZbin free compare with that course is their any variation??

[@bc090200352]

Reject legitimate email as spam

[@rajkir2852]

False +ve : when my doctor says i am covid positive but i was negative. Me to doctor: give my money back. I need that money for buying cigarateee

[@dhanushappala8685]

😂😂

[@saiteja3437]

That is false positive

[@bhavesh3913]

Description telegram link not working...please look on it

[@sashankmvv4335]

Op

[@ArunKumar-fv6uw]

In first case, I believe the model should reduce both false negative and false positive. Yup I agree that false negative cannot be tolerated. At the same time, false positive too has an adverse impact on health. In false positive case, a person should end up with spending money for further examination/treatment(s) even though actually it is not necessary. such unnecessary treatments can cause risks. In the case of covid like diseases, a model must equally avoid false positive also as covid hospitalisation can easily drain large sum of money.

[@datavisuals3276]

the train will be late today

[@bishwajeetsingh8834]

Spam mails

[@erkamaalhussain5881]

Fake news classifier