-
Regression vs Classification:
Regression Classification Predicted values Continuous Discrete Intuition Predictions are points on a hyper-plane. Predictions are points in hyper-volumes divided by a hyper-plane Linear vs Polynomial Hyper-plane vs Hyper-surface Hyper-volumes on 'either side' of Hyper-plane vs Hyper-surface E.g. for $n = 1$ , output isStraight line in $\mathbb{R}^2$ Areas on either side of a straight line in $\mathbb{R}^2$ Linear vs Polynomial Hyper-plane vs Curved Hyper-surface in $\mathbb{R}^{n+1}$ Hyper-volumes on 'either side' of Hyper-plane vs Curved Hyper-surface in in $\mathbb {R} ^{n+1}$ -
Octave cheatsheet:
Command Action pinv/inv matrix (psuedo)inverse format long/short Precision of double display sprintf/printf/disp Format strings/Display value 1.0:0.1:2.9 Range zeros/ones/rand/eye(r, c) Commonly used matrices size, length Dimensions of matrix who/whos List objects save/load serialization of data A(r, c)/A(2, :)/A([1 4], :) read/write cells/rows/columns [A, [...]] append row/columns A(:) unroll [A B]/[A; B] Concat column/row wise /. matrix multiply/matrix element-wise multiply [val ind] = max/max Value/index of max elementwise compare op find(compare op) sum/prod/floor/ceil flip/fliplr/flipup Flip stuff plot hold on/off Reuse/Create new figure clear close print -dpng print the plot in a bmp figure select figure subplot picture - in - picture axis change axis imagesc visualize matrix as grid of colors colorbar, colormap , chain commands ; dont print command output timeit/tick-tock Time steps
parthopdas / c-machine-learning Goto Github PK
View Code? Open in Web Editor NEW[Coursera] Machine Learning
Home Page: https://www.coursera.org/learn/machine-learning