in graph of life satisfactation vs gdp if distribution of instances is complex means not linearly increasing like kind of circular. so can we say by looking at graph that linear regression model will not be effective model to predict life satisfaction?
If the data is not linearly separable, we usually do not use LinearRegression but some other model, or we use a model of higher complexity, depending on the distribution.
Precision is the number of positive identifications that are actually correct. Whereas, recall is the proportion of actual positives that were identified correctly. Also, it is not a shopfilter but a shoplifter, it is a person who steals from shops.
Never5Classifier is a fake classifier or toy classifier, which doesn't learn anything but just returns zeros(or False) no matter what the input features are.
The part of the video you just referred to, that is not because of the video. That was because Sandeep was finishing a sentence and so his voice went down. However, thank you for your feedback.
1) What is the significance of using 2 in the code (X = 2 * np.random.rand(100, 1))? If we use other values what will it impact and how.
2) What is "b." in plt.plot(X, y, "b.") + Why did we choose the scale as 0->2 for x-axis and 0->15 for y-axis (plt.axis([0, 2, 0, 15]))? Any specific reasons?
Thanks
___________________________________________
import numpy as np
# Uniformly distributed array with dimension (100, 1)
X = 2 * np.random.rand(100, 1)
# Normally distributed array with noise
y = 4 + 3 * X + np.random.randn(100, 1)
1. Uniform distribution is a type of probability distribution where all outcomes are equally likely. A coin has an uniform distribution because the probability of getting either heads or tails in a coin toss is the same. A normal distribution, also called the Gaussian distribution, is a probability distribution commonly used to model phenomena such as physical characteristics (e.g. height, weight, etc.) and test scores. For normal distribution the mean, mode and median are all equal. The curve is symmetric at the center (i.e. around the mean, μ). Exactly half of the values are to the left of center and exactly half the values are to the right. The total area under the curve is 1.
2. We are creating this distribution so that it gives us a dataset on which we can perform linear regression.
Hi. I couldn't find training_linear_models.ipynb from My Lab. Is there another specific path? Please note I have skipped many contents in the middle for now, to start some hands on experience. I would like to execute some examples and learn parallelly.
I would suggest you to clear your browser cache and cookies, if possible use a different browser. Also, check if your firewall is preventing from accessing this site.
Your notebook is working fine, i can see pages in other tabs are also unresponsive so i think there might be a issue with the internet i would request you to check with a different internet connection.
A decent bandwidth is required to use mutiple notebooks simultaneously.
If this doesn't help you please write us at reachus@cloudxlab.com.
The ROC AUC score for SGD classifier is 0.9637386758107422
The ROC AUC score for Random Classifer is 0.995175585758131
Also,in your example of y_train_5 :
The ROC AUC value for SGDClassifier : 0.9597417516899552
The ROC AUC value for Random Forest Classifier is : 0.9983730844244434
As in both the example we can clearly see that Random Classifer performs much better than SGD classifer,so we will be going with the Random Classifer model,so
1: What is the importance of SGD classifer over the Random Classifer?
2: When we know that Random classifer will give more good results than SGD classifer,then why we have to spent time on SGD classifer?
What is the most common model/ algorithm used for image classification?
What is the most common algorithm used for natural language understanding and generation?
what are all the different python models used in machine learning, deep learning, AI? Other languages use the same models/algorithms in their language?
what is the difference between machine learning, deep learning, AI?
There are no commong models. The choice of models depends on various factors like dataset, data type, the problem you are trying to resolve etc. Also, for a detailed discussion on difference between ML, DL, and AI, please go through the course lectures.
What is the most common model/ algorithm used for image classification?what are all the different python models used in machine learning, deep learning, AI, Image classification? Other languages use the same models/algorithms in their language?
What is the most common algorithm used for natural language understanding and generation?
As I said in my previous comment, there are no common models. The choice of models depends on various factors like dataset, data type, the problem you are trying to resolve etc. There is no one-model-fits-all theory when it comes to Machine Learning or Deep Learning.
Hello, Role of Bias is only to shift the best fit line from origin. i am not getting the role of Bias. . will the model classify correctly if we don't use Bias..
Bias is like intercept in the lines, which cuts the y axis. When we compare an ideal model to the real time model which may not be exact and there is some error in the actual and predicted, this is called bias. or Bias is the inability of the Machine/ data to capture the true relationships. b/t actual and predicted.
Suppose there is no inputs (xi) but in some cases we need some output as a base, to make our model work or else we will limit the expressivity of our linear model. That is why we use some initial bias. This is also a parameter which is learnt. Initially to start take bias as 1.
In session 12 at 2:54:59 the plot after changing y of the example to z the line of regression obtained is wrong (I was also getting this error). This was because of theta_best of the example replaced the correct value of theta best Changing theta_best of example to best_theta will solve the problem.
@sandeepgiri: After going through this session, I wish you were my maths professor during my college days. Then, instead of simply mugging my text, I would have understood why we were doing something. Thanks for making the maths in ML such fun.
Please login to comment
47 Comments
I cannot find the training_linear_models notebook in my lab.
Upvote ShareHi,
1:29:45 - Can you please explain how we came up with the bias -
# Add Bias term x0 = 1
X_b = np.c_[np.ones((100, 1)), X]
How do we decide what bias needs to be added and when do we add it?
Upvote ShareWe just choose any initial random value, the same way we initialize weights. You will learn it in more details in further videos.
Upvote ShareDear CloudX Lab Team,
Checking the course contents once again after a few months gap.
Am unable to find training_linear_models.ipynb. Could you guide/help me regarding this.
Thanks & Regards,
Upvote ShareHi;
in graph of life satisfactation vs gdp if distribution of instances is complex means not linearly increasing like kind of circular. so can we say by looking at graph that linear regression model will not be effective model to predict life satisfaction?
Upvote ShareHi,
If the data is not linearly separable, we usually do not use LinearRegression but some other model, or we use a model of higher complexity, depending on the distribution.
Thanks.
Upvote ShareI did not understand the video at 1:48:50
Precision and recall
How can we use precision for unsafe videos and recall for shop filter??
What is shoapfilter??
Upvote ShareHi,
Precision is the number of positive identifications that are actually correct. Whereas, recall is the proportion of actual positives that were identified correctly. Also, it is not a shopfilter but a shoplifter, it is a person who steals from shops.
Thanks.
Upvote Sharewhat is Never5classifier??
Is it a classifier like sgd or it is a program??
Upvote ShareHi,
Never5Classifier is a fake classifier or toy classifier, which doesn't learn anything but just returns zeros(or False) no matter what the input features are.
Thanks.
Upvote Sharehi,
there's inconsistency in audio rendering like the loudness it ranges from medium level to too low
Upvote ShareHi,
Thank you for your feedback. We would consider these and try to rectify the same when we are updating our course lecture videos.
Thanks.
Upvote Shareeg : @58:38 just watch 3-4 sec voice goes down very low
Upvote ShareHi,
The part of the video you just referred to, that is not because of the video. That was because Sandeep was finishing a sentence and so his voice went down. However, thank you for your feedback.
Thanks.
Upvote ShareHello,
I have 2 questions from the code below:
1) What is the significance of using 2 in the code (X = 2 * np.random.rand(100, 1))? If we use other values what will it impact and how.
2) What is "b." in plt.plot(X, y, "b.") + Why did we choose the scale as 0->2 for x-axis and 0->15 for y-axis (plt.axis([0, 2, 0, 15]))? Any specific reasons?
Thanks
___________________________________________
import numpy as np
# Uniformly distributed array with dimension (100, 1)
X = 2 * np.random.rand(100, 1)
# Normally distributed array with noise
y = 4 + 3 * X + np.random.randn(100, 1)
plt.plot(X, y, "b.")
Upvote Shareplt.xlabel("$x_1$", fontsize=18)
plt.ylabel("$y$", rotation=0, fontsize=18)
plt.axis([0, 2, 0, 15])
plt.show()
Hi,
1. You can choose any number instead of 2 here. We are simply generating a data which is linear in nature.
2. We have chosen these axes values because the data fits within that.
Thanks.
Upvote ShareHi. In slide 53, we have generated some linear looking data. My doubt is:
1) What is uniformly distributed array and normally distributed array, and what's the difference between them?
2) Why are we using them to create linear data? I cant understand the Mathematical logic, kindly explain.
1 Upvote ShareHi,
Good question!
1. Uniform distribution is a type of probability distribution where all outcomes are equally likely. A coin has an uniform distribution because the probability of getting either heads or tails in a coin toss is the same. A normal distribution, also called the Gaussian distribution, is a probability distribution commonly used to model phenomena such as physical characteristics (e.g. height, weight, etc.) and test scores. For normal distribution the mean, mode and median are all equal. The curve is symmetric at the center (i.e. around the mean, μ). Exactly half of the values are to the left of center and exactly half the values are to the right. The total area under the curve is 1.
2. We are creating this distribution so that it gives us a dataset on which we can perform linear regression.
Thanks.
1 Upvote ShareHi. I couldn't find training_linear_models.ipynb from My Lab. Is there another specific path? Please note I have skipped many contents in the middle for now, to start some hands on experience. I would like to execute some examples and learn parallelly.
Upvote ShareHi,
Please find below the path to the notebook:
https://github.com/cloudxlab/ml/blob/master/machine_learning/training_linear_models.ipynb
Thanks.
Upvote Sharein this the input number [47]
this code can be replaced by grid search, mean squared error(as shown in bike rental forecasting project?
Also this project .ipynb is not in files section to download , please provide
Upvote Sharescreenshot:
Hi,
Could you please tell me which project are you referring to since this is the Training Models chapter.
Thanks.
Upvote Sharehello,
I am unable to access training_linear_models.ipynb.
My page hangs trying to restart kernal. Finally gives error-
Cannot open. This page is having problem.
kindly help
Sumbul
Image attatched...
Upvote ShareHi,
It is working fine on my end, could you please let me know if you are stil facing this issue.
Thanks.
Upvote ShareYes, it still the same isue.
Not openining same error..
Upvote ShareHi,
I would suggest you to clear your browser cache and cookies, if possible use a different browser. Also, check if your firewall is preventing from accessing this site.
Thanks.
Upvote ShareTried clearing history n cache too still not working.
Opens the page for few sec.. Shows wait kernal getting ready...
Then hangs with above error..
Frustrating...
Upvote ShareHi Sumbul,
Your notebook is working fine, i can see pages in other tabs are also unresponsive so i think there might be a issue with the internet i would request you to check with a different internet connection.
A decent bandwidth is required to use mutiple notebooks simultaneously.
If this doesn't help you please write us at reachus@cloudxlab.com.
All the Best!
Upvote Sharesachin the issue is not with internet at my end..
just checked the speed etc
kindly see the attached image
Upvote ShareHi Sumbul,
I have replied to you email, Please check.
Upvote ShareHello,
My ROC Auc for y_train_9:
Also,in your example of y_train_5 :
As in both the example we can clearly see that Random Classifer performs much better than SGD classifer,so we will be going with the Random Classifer model,so
1: What is the importance of SGD classifer over the Random Classifer?
2: When we know that Random classifer will give more good results than SGD classifer,then why we have to spent time on SGD classifer?
Please clear my both the points.
Thanks.
Upvote ShareHi,
There is no one rule that fits all situation in Machine Learning or Deep Learning. In some datasets SGDClassfier might give better results than RF.
Thanks.
Upvote ShareOk sir,Thank-You for your answer.
Upvote ShareWhat is the most common model/ algorithm used for image classification?
What is the most common algorithm used for natural language understanding and generation?
what are all the different python models used in machine learning, deep learning, AI? Other languages use the same models/algorithms in their language?
what is the difference between machine learning, deep learning, AI?
Upvote ShareHi,
There are no commong models. The choice of models depends on various factors like dataset, data type, the problem you are trying to resolve etc. Also, for a detailed discussion on difference between ML, DL, and AI, please go through the course lectures.
Thanks.
Upvote ShareThanks
What is the most common model/ algorithm used for image classification?what are all the different python models used in machine learning, deep learning, AI, Image classification? Other languages use the same models/algorithms in their language?
What is the most common algorithm used for natural language understanding and generation?
Upvote ShareHi,
As I said in my previous comment, there are no common models. The choice of models depends on various factors like dataset, data type, the problem you are trying to resolve etc. There is no one-model-fits-all theory when it comes to Machine Learning or Deep Learning.
Thanks.
Upvote Sharehow can i download the ppt
Upvote ShareHi,
You can find the option to download the slides on the bottom left of the slide content.
I hope this helps if you face any further issues feel free to contact us again. Happy to help.
Thanks.
-- Mayank Sharma
Upvote ShareHello,
Upvote ShareRole of Bias is only to shift the best fit line from origin. i am not getting the role of Bias. . will the model classify correctly if we don't use Bias..
Hi Kushboo.
Bias is like intercept in the lines, which cuts the y axis. When we compare an ideal model to the real time model which may not be exact and there is some error in the actual and predicted, this is called bias.
or
Bias is the inability of the Machine/ data to capture the true relationships. b/t actual and predicted.
All the best!
-- Satyajit Das
Upvote Shareok so for that we have to take bias =1.. if not then what factor bias value depends?
Upvote Sharethank you for quick response
Good question!
Suppose there is no inputs (xi) but in some cases we need some output as a base, to make our model work or else we will limit the expressivity of our linear model. That is why we use some initial bias. This is also a parameter which is learnt. Initially to start take bias as 1.
All the best!
-- Satyajit Das
Upvote ShareIn session 12 at 2:54:59 the plot after changing y of the example to z the line of regression obtained is wrong (I was also getting this error).
Upvote ShareThis was because of theta_best of the example replaced the correct value of theta best
Changing theta_best of example to best_theta will solve the problem.
@sandeepgiri: After going through this session, I wish you were my maths professor during my college days. Then, instead of simply mugging my text, I would have understood why we were doing something. Thanks for making the maths in ML such fun.
Upvote Share@Sundeep may I ask you why this gradient matrix at video 2:49:28 needs X Transpose product
thank you
Upvote Share