ML Andrew Ng

Question 1

what is the diff btwn function with single feautures and one with multipple feautures

Answer 1

Question 2

How to drive Multi variate formual with Transpose and X

Answer 2

Question 3

What is parameter ?

Answer 3

Theta is the parameter

Question 4

How to fit the parameter (to obtain best one) in the hypothesis (prediction equation)

Answer 4

Using gradient descend

Question 5

What is the cost function for both linear and mulytivariate feature

Answer 5

Question 6

Gradient descend formula for both one variable and double

Answer 6

Question 7

If you have a problem with multiple feautures , you should make sure those feautures have similar scales ….therefore

Answer 7

gradient descend will converge more easily

Question 8

example of future scaling ranges

Answer 8

Question 9

Gradient descent feauture scaling and normalization

Answer 9

Question 10

example of feauture scaling

Answer 10

Question 11

What is the job of gradient descent?

Answer 11

The job of gradient descent is to find the value of theta for you that hopefully minimizes the cost function J(theta).

Question 12

How to know if gradient descent is working?

Answer 12

If the gradient descend is workng properly , then J(Q)

Should decreas after every iteration . If Gdecent is increasing , it means , it is not working

Question 13

cheking gradient descent is working

Answer 13

Question 14

Gradient Descent Learning Rate

Answer 14

Question 15

To choose alpha or learning rate , use ?

Answer 15

0.001 , 0.01, 0.1,1… or u can go for 3 fold increase , 0.001 , 0.003,0.009..

Question 16

what will happen if alpha is too small?

Answer 16

then the learning rate is will be slow

Question 17

polynomial regression?

Answer 17

Question 18

feautues and polynomial regression

Answer 18

Question 19

What is normal equation

Answer 19

Question 20

“What does the machine (? what is the machine here) actually learn?”

Answer 20

i.e. the statistical model

Question 21

……Grew from the field of AI.

Answer 21

Machine Learning

Question 22

Examples of where machine learning can be applied?

Answer 22

Database mining , application that we cannot program by hand( NLP , autonomous heli , computer vision) ,sel-customizing program(Netfliz , Amazon ,) , understanding human brain

Question 23

not well defined , but a field of study that gives computers the ability to learn without without being explicitly programmed

Answer 23

Definition of ML , by Samuel

Question 24

Well posed learning probelm , A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, it is performance at tasks in T, as measured by P , improves the expereince E.

Answer 24

Tom Michel. Example samule wrote a checkers playing program..had the program play 1000o games against itslef. E= 10000 examples , T is playing checkers , P if you win or not

Question 25

Supervised learming

Answer 25

Teach the computer how to do something , then let it use its new found knowldege to do it

Question 26

Unsupervise

Answer 26

Let the computer learn how to do something ,and use this to determine the structure and patterns in data

Question 27

Other learning , we have reinforcement learning and recommender system

Answer 27

part of machine learning tech

Question 28

The most common type of probelm in machine learning

Answer 28

Probably the most common probelm type in machine learning

Question 29

Predicting continue value output is called?

Answer 29

Regression probelm ?

Question 30

Classifying data into 2 discrete classes - no in between

Answer 30

is called Classification probelm.

Question 31

Support-vector machine support infinite number of feautures

Answer 31

one example

Question 32

which learning algorihm is using unlabbled data?

Answer 32

unsupervise learning algoritm

Question 33

just told , here is the data , can you structur it?

Answer 33

one way of doing this is cluster analysis......which is unsupervise learning.

Question 34

Examples of clustering ?

Answer 34

organizing news , genomics , organize computer clusters , social network analysis (customer data) ,

Question 35

input variables of feautures and output variable or target

Answer 35

input an doutput in Rgression

Question 36

In the case of the simple ...... (y approximately equal b0 + b1 \* X where X is one column/variable) the model “learns” (read: estimates) two parameters; b0: the bias (or more traditionally the intercept); and, b1: the slope

Answer 36

Simple Linear Regression

Question 37

The bias is the level of y when X is 0 (i.e. the value of sales when advertising spend is 0) and the slope is the rate of predicted increase or decrease in y for each unit increase in X (i.e. how much do sales increase per pound spent on advertising). Both parameters are scalars (single values).

Answer 37

Regression parameters

Question 38

Univariate Linear Regression is also called....

Answer 38

Linear Regression (with one parameter)

Question 39

What is the form of Linear Regression?

Answer 39

Question 40

What is cost function?

Answer 40

A cost function let us find how best to fit straigght line on our data

Question 41

Choosing values for Theta gives differnt ?????

Answer 41

functionand threfore diff regression line

Question 42

Base on our training set , we want generate parameters that .....

Answer 42

makes straight line. To fromalizw this we are trying to solve minimization problem. that is minimize diff btw h(x)-y for every every and each example and sum it over training set

Question 43

In ML, ........are used to estimate how badly models are performing. Put simply, a cost function is a measure of how wrong the model is in terms of its ability to estimate the relationship between X and y. This cost function (you may also see this referred to as loss or error.) can be estimated by iteratively running the model to compare estimated predictions against “ground truth” — the known values of y. This is typically expressed as a difference or distance between the predicted value and the actual value. The cost function (you may also see this referred to as loss or error.) can be estimated by iteratively running the model to compare estimated predictions against “ground truth” — the known values of y. The objective of a ML model, therefore, is to find parameters, weights or a structure that minimises the cost function.

Answer 43

cost function. also called lost function.

Question 44

cost function details formula

Answer 44

Question 45

......is like your prediction machine , throw it X and will get putative y

Answer 45

Hypothesis. We want summarize cost function. and because of summation terms , it is ineherently looking at all the data in the training set at any time.

Question 46

cost function

Answer 46

Question 47

cost function intuition calculation

Answer 47

Question 48

The function is called square error cost function is ,.....

Answer 48

cost function

Question 49

...............is a way to using ur tarining data , determine values for your theta which makes the hypothesis as accurate as possible.

Answer 49

cost or cost function

Question 50

Most common use function for regression iss....

Answer 50

cost function

Question 51

cost function general representation

Answer 51

Question 52

What is cost function objectives?

Answer 52

Question 53

Gradient descent

Answer 53

So we have our hypothesis function and we have a way of measuring how well it fits into the data. Now we need to estimate the parameters in the hypothesis function. That's where gradient descent comes in. Gradient descent minimize cost function

Question 54

Once the model learns these ........s they can be used to compute estimated values of y given new values of X. In other words, you can use these learned parameters to predict values of y when you don’t know what y is — hey presto, a predictive model!

Answer 54

parameters ....In Linear regression

Question 55

Learning parameters: Cost functions There are several ways to learn the parameters of a LR model, I will focus on the approach that best illustrates statistical learning;........

Answer 55

minimising a cost function.

Question 56

Remember that in ML, the focus is on learning from data. This is perhaps better illustrated using a simple analogy. As children we typically learn what is “right” or “good” behaviour by being told NOT to do things or being punished for having done something we shouldn’t. For example, you can imagine a four year-old sitting by a fire to keep warm, but not knowing the danger of fire, she puts her finger into it and gets burned. The next time she sits by the fire, she doesn’t get burned, but she sits too close, gets too hot and has to move away. The third time she sits by the fire she finds the distance that keeps her warm without exposing her to any danger. In other words, through experience and feedback (getting burned, then getting too hot) the kid learns the optimal distance to sit from the fire. The heat from the fire in this example acts as a ........ — it helps the learner to correct / change behaviour to minimize mistakes.

Answer 56

cost function

Question 57

In ML, cost functions are used to estimate how badly models are performing. Put simply, a cost function is a measure of how wrong the model is in terms of its ability to estimate the relationship between X and y. This is typically expressed as a difference or distance between the predicted value and the actual value. The cost function (you may also see this referred to as loss or error.) can be estimated by iteratively running the model to compare estimated predictions against “ground truth” — the known values of y. The objective of a ML model, therefore, is to find parameters, weights or a structure that minimises the cost function.

Answer 57

ML

Question 58

Now that we know that models learn by minimizing a cost function, you may naturally wonder how the cost function is minimized — enter ,,,,,,,,. Gradient descent. it is an efficient optimization algorithm that attempts to find a local or global minima of a function.

Answer 58

gradient descent. Gradient descent enables a model to learn the gradient or direction that the model should take in order to reduce errors (differences between actual y and predicted y). Direction in the simple linear regression example refers to how the model parameters b0 and b1 should be tweaked or corrected to further reduce the cost function. As the model iterates, it gradually converges towards a minimum where further tweaks to the parameters produce little or zero changes in the loss — also referred to as convergence.

Question 59

This process is integral (no calculus pun intended!) to the ML process, because it greatly expedites the learning process — you can think of it as a means of receiving corrective feedback on how to improve upon your previous performance. The alternative to the gradient descent process would be brute forcing a potentially infinite combination of parameters until the set that minimizes the cost are identified. For obvious reasons this isn’t really feasible. Gradient descent, therefore, enables the learning process to make corrective updates to the learned estimates that move the model toward an optimal combination of parameters.....which process?

Answer 59

Gradient Descent.

Question 60

......minimise cost function and , use all over place for ML

Answer 60

Gradient decent

Question 61

How gradient descent works?

Answer 61

Question 62

Gradient Descent Algorithm

Answer 62

Question 63

What is alpha ???

Answer 63

Alpha is a learning rate . it control how big steps u take. .if alpha is big , u have aggreesive gradient desent, if alpha is small take tiny steps for gradient descent

Question 64

How Gradeint Descent is implemented?

Answer 64

Question 65

understdaning gradient descent

Answer 65

Question 66

Linear Regression with gradient descent

Answer 66

Question 67

There are other method for finding solution to minimum function...what are they ?

Answer 67

Normal equation method , but gradient descent scale better to large dataset and use in alot of contxt and machine learning

Question 68

Normal equation

Answer 68