Linear Regression 1 [Python]

[@CausticTitan]

I'm still trying to wrap my head around the camera/display setup that allows him to write in front of himself but keeps it from looking backwards for us
It really adds to the lecture!

[@heinsaar]

There is no magical setup. It's done by flipping the original video (where everything _is_ backwards for us) along the vertical center.

[@navinbondade5365]

Leo Heinsaar can you please explain in detail

[@SkielCast]

@Tech Axis and @CausticTitan
This is done following these steps:
1. Use a transparent board compatible with markers.
2. Put the camera behind it.
3. Use a really dark background and also make the host wear dark clothes.
4. Since the camera was behind the board, the recording will be inverted, so it should be mirrored (vertically inverted).
5. Make some color adjustments so that the background and clothes dissapear without corrupting what's written.

[@vivekkumbhar3809]

Thanks for this Video. I am just starting with Machine Learning and this helped me understanding Linear Regression along with a pretty example. Also a very nice display Setup.

[@PaulNielan]

For those experimenting in Mathematica, try these commands---(from PEN)
x = 3
delta = .25
a = Range[-2, 2, delta]
b = x*a + 1* RandomReal[{-1, 1}, Length[a]]
p1 = ListPlot[Transpose[{a, b}], PlotStyle -> Red,
PlotLegends -> {"Raw Data"}]
p2 = ListLinePlot[{Transpose[{a, x*a}]}, PlotStyle -> Blue,
PlotLegends -> {"Exact"}]
Show[p1, p2]
amat = Transpose[{a}]
{U, S, V} = SingularValueDecomposition[amat, MatrixRank[amat]]
xtilde = V . Inverse[S] . Transpose[U] . b // First
p3 = ListLinePlot[{Transpose[{a, xtilde*a}]}, PlotStyle -> Green,
PlotLegends -> {"Regression"}]
Show[p1, p2, p3]

[@GauravSharma-ui4yd]

Do normalizing the data and then applying SVD really make change? I mean does it helps normalizing data before applying SVD?

[@alial-musawi9898]

No. Ordinary Least Squares is scale-invariant.
However, for Ridge Regression, you should normalize.

[@erockromulan9329]

I understand the reason behind only using some of the data to verify your model, but would it still be 'cheating' to use all of the data to develop a more robust model to predict other homes outside of the data set?

[@threedworld2319]

where is the jupyter notebook? cannot find it on website databookuw.com

[@nskmodnar]

pip install jupyter

[@PetersonCharlesMONSTAH]

Hey bro, can you help me out with from sklearn.linear_model import LinearRegression? I get an error message from using.
This is the error message I'm getting. from R and python.ModuleNotFoundError Traceback (most recent call last)
in ()
----> 1 from sklearn.Linear_model import LinearRegression
2 regressor = LinearRegression()
3 regressor.fit(X_train, y_train)
ModuleNotFoundError: No module named 'sklearn.Linear_model'
Thanks bro.

[@alial-musawi9898]

'Linear_model' should be 'linear_model'.

[@Alexjohnson7929-p2n]

Color, LineWidth, MarkerSize
these should all be in lower case in the new version of matplotlib

[@dr.gordontaub1702]

THANK YOU. This is the comment I was looking for to find out why my version of the code wasn't working.

[@apoorvshrivastava3544]

good morning sir