Live-Discussing All Hyperparameter Tuning Techniques Data Science Machine Learning

[@harikrishna-harrypth]

A TRUE LEGEND AND MASTER OF DATA SCIENCE!!!! THANK YOU KRISH NAIK!!! YOU'RE A REAL GEM FOR THE WORLD OF DATA SCIENCE!!!!! GOD BLESS YOU MAN! ✌️💖

[@jesuskristus18]

Great, another Indian/Pakistani “data scientist” from Fiverr.

[@rohitbharti2882]

Explained so well. My confidence in DS increases day by day through your videos. 😊

[@Gester2000]

Let me tell u you are the gem of the game out of all the ones teaching data science on KZbin passing us real world thought process of a datascientist working in a real world scenarios Love from Karachi Pakistan,🇵🇰

[@jayanthAILab]

I love ur clarity on the subject . Best teacher in the youtube

[@Schneeirbisify]

great work, very clear and helpful for my project that I am working on. Thanks a lot!

[@vijaisrivastava1629]

krish bhai... thank you for your help .. i know 20 data scientist who made transition because of your help.. jai sree ram..

[@natarajanlalgudi]

Thanks again Krish Naik amazing efforts and commitment so grateful.

[@amrutabagalkot6407]

BEST VIDEO EVER .....HATS OF TO YOU SIR 🙏

[@priyanshshankhdhar347]

please do a video on FasterRCNN and Yolo object detection

[@srinivasarukonda8768]

Krish really amazing knowledge sharing ..gr8 work..

[@jaiganeshnagidi5716]

Sir please make a video on yolo object detection 🙏

[@kartikjswl4]

Damn!!! I couldn't thank you enough for this ever.. 🙏🏻🙏🏻

[@soulrider6822]

Hi Krishna I have seen your most of the video

[@priyanshshankhdhar347]

if possible.. please do video on faster rcnn and yolo object detection without github repo.. or even with github repo.

[@ahmeterdonmez9195]

1:05 question from watchers : "why multiplication?". The result of embarking on this path without learning the necessary mathematics.... So it's a simple probability case. God help them in situations that require more advanced math.

[@mutyaluamballa]

U just covered all the stuff in one cool video, this just blew my mind bro. I just cant say one reason for not subscribing your channel. Thank you very much...! 💕

[@kumawatrohan]

Thankyou so much sir for this detailed explanation ❤️

[@rayyanmohsin8638]

Shouldn't we split our data before imputing any sort of values to prevent data leakage?

[@sunilabans1]

Thanks for sharing the knowledge.

[@write2ruby]

9:30 GridSearchCV and RandomizedSearchCV are good

[@vaibhavshukla0729]

Thank you sir 🌟

[@ajaykushwaha-je6mw]

first time I ma seeing you in funny mood, good to see you like this else aap to bhagwan shanker ki tere gusse mein hi dikhte hain.

[@rayyanmohsin8638]

Starts at 11:30

[@fozler]

Hello sir, I am from Bangladesh and always watch your video. Can you make some videos about fusion models.

[@vargabghosh5497]

But using tpot can I print the values of the hyper parameter for which our model has best accuracy...

[@justthink8319]

THANK YOU BRO IT WAS AMAZING SESSION

[@jiyabyju565]

thank you sir...why dont i get accuracy value..? so there is no return value on loss

[@sandipansarkar9211]

finished watching

[@tusharpatil1957]

Telegram link is not opening

[@mikefda12]

hey question how do you get the predictive text?

[@manassrivastava6452]

WHEN WILL OBJECT DETECTION GOING LIVE ??

[@ashishsaini5096]

how u r not getting error while u having 1 as int value in min_samples_split which is not allowed ! although i m getting this error (min_samples_split must be an integer greater than 1 or a float in (0.0, 1.0]; got the integer 1) which is right : we can either use 1.0 float or greater value than int 1

[@dhirendrakumarjha7385]

can i implement these concept if i have continuous value as output ie if I want to do regression problem

[@vidyamc4340]

Grid search will be best I guess

[@sajidchoudhary1165]

Sir Please makes video on Mathematics behind on SVM Regression, AdaBoost Regression, Gradient Boost Classification

[@aparnarout2008]

Good evening sir, I needed some guidance how can I can contact with you?

[@harshmalviya7]

I have tried hyper parameter and my laptop take 6 hrs to give the parameter what should I do ! It is wasting my time.

[@LikithVibes]

You can run the same code in kaggle.Kaggle provides free access to NVidia K80 GPUs in kernels

[@shootgold]

Try google colab

[@shahnawazkhan1636]

Best session please conduct such kind of class

[@pradheepm1371]

How to reduce the false positive and false negative

[@LikithVibes]

As per my understanding and knowledge, if your data is balanced in terms of proportion of two classes and if you have built very good model then automatically your false positive and false negative will be less. But if your data is imbalance, depending on the use case that you are working on you can increase or decrease the threshold to reduce false positive and false negative. But that's a tedious process , so better way is to look at ROC curve.

[@neetikagupta8536]

can we do stratifiedkfold validation in gridsearchcv or randomsearchcv

[@sahilp4796]

Yes, we can use. Sending you a sample code for RandomizedSearchCV skf = StratifiedKFold(n_splits = 5, shuffle = True, random_state = 7) random_search = RandomizedSearchCV(model, param_distributions=params, n_iter=3, scoring='accuracy', n_jobs = -1, cv = skf.split(X_train, y_train), random_state=7)

[@hiteshyerekar2204]

HIi Krish I got this error how to solved it. ValueError: Invalid parameter min_sample_split for estimator RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None, criterion='gini', max_depth=None, max_features='auto', max_leaf_nodes=None, max_samples=None, min_impurity_decrease=0.0, min_impurity_split=None, min_samples_leaf=1, min_samples_split=2, min_weight_fraction_leaf=0.0, n_estimators=250, n_jobs=None, oob_score=False, random_state=None, verbose=0, warm_start=False). Check the list of available parameters with `estimator.get_params().keys()`.

[@ai_beyond_boundaries]

i also got the same error

[@bharadwajnarayanam9922]

Hi Hitesh! Can you show the code too?

[@hiteshyerekar2204]

@@bharadwajnarayanam9922 hiii I solved those problem.

[@bharadwajnarayanam9922]

@@hiteshyerekar2204 Cool bro!

[@anielkali704]

Krish you are amazing, keep it up! One comment, I wouldn't take too high values ​​for the 'max_depth' parameter because of overfitting issues...

[@akarshankumar1711]

It's okay to take high values anyways it's random forest, a high variance base model is needed. And also it's precisely not depth but more related to num of leafs. Hence high value do more good than harm.

[@CreatingUtopia]

I got an error :cant pickle file and send it to workers when i ran the randomsearch cv

[@its_me7363]

remove 'n_jobs' parameter.

[@amexethiotech1619]

hi good evening

[@amexethiotech1619]

yes

[@sunilabans1]

Yes

[@lemuelkbj493]

7:33

[@parthagarwal4592]

When you are pissed off of copy pasting things - 46:52

[@sandipansarkar9211]

finished coding

[@sunilabans1]

Yed

[@sunitabnsl]

the lecture is good but shaking legs does not seem good kris.

[@AchinAbhi]

Hello everyone, I get an error regarding accessing subscript for the randomizedsearchcv object, 1 from sklearn.model_selection import GridSearchCV 2 param_grid = { ----> 3 'criterion': [rf_randomcv['criterion']], 4 'max_depth': [rf_randomcv['max_depth']], 5 'max_features': [rf_randomcv['max_features']], TypeError: 'RandomizedSearchCV' object is not subscriptable

[@pragatiagrawal3599]

Thank you so much sir 😊😊

[@karthiksundaram544]

Yes