Support Vector Machine + AdaBoost + Linear Regression + Logistic Regression + K-Nearest Neighbors + Math + Python

Download source code @ https://sites.fastspring.com/prototypeprj/instant/ai

• SVM + SMO + Python (Prototype Project 01)

  • demo a prebuilt version of the application

  • code the application

  • go over the training data used in this app.

  • go over the various classes that make up the app.

  • quick introduction to Support Vector Machine (SVM)

  • quick introduction to Sequential Minimal Optimization (SMO)

  • code the SupportVectorMachines class

  • plug in equation for 'w' into equation of a linear svm and use resulting equation in code

  • EPSILON and slack penalty C

  • alpha and violating KKT conditions

  • selecting the index of the 2nd alpha to optimize

  • optimize an alpha pair and b

  • define and calculate w

  • classify method

  • display information tables code

  • handle command line entry code

  • plot data + decision boundary + support vectors code

  • explain + test run application

  • change C and rerun app. to show overfitting

• AdaBoost + Python (Prototype Project 01)

  • demo a prebuilt version of the application

  • code the application

  • go over the training data used in this application

  • go over the various classes making the application

  • go over boosting

  • go over data classification with AdaBoost

  • explain + code the TreeTrunk classifier class

  • explain + code the AdaBoost class

  • explain + code the DisplayHelper class

  • explain + code the handle_command_line entry function

  • code the application

  • test run the application

• Linear Regression + Python + Normal Equation (Prototype Project 01)

  • demo a prebuilt version of the application

  • code the application

  • go over the training data used in this app.

  • use matplotlib to plot the data

  • linear regression hypothesis and normal equation

  • implement normal equation

  • throw exception if determinant does not have an inverse

  • estimate rent method

  • handle command line entry function

  • test run completed app.

• Logistic Regression + Gradient Descent + Python (Prototype Project 01)

  • demo a prebuilt version of the application

  • code the application

  • training data used in this app.

  • logistic regression hypothesis

  • logistic/sigmoid function

  • gradient of the cost function

  • update weights with gradient descent

  • implement logistic method in LogisticRegression class

  • implement gradient_descent method in LogisticRegression class

  • implement classify method in LogisticRegression class

  • function using matplotlib for plotting candidates scores and decision boundary

  • command line entry function

  • test run the completed app.

• Logistic Regression + Stochastic Gradient Descent + Python (Prototype Proj 02)

  • demo a prebuilt version of the application

  • code the app.

  • go over training data used in this app.

  • gradient descent

  • stochastic gradient descent

  • LogisticRegression class from previous tutorial modified to include stochastic gradient descent implementation

  • create a new Logisticregression instance and call the stochastic gradient descent method on it

  • plot function implementation using matplotlib

  • handle command line entry function

  • test run completed app.

• K-Nearest Neighbors + Python (Prototype Project 01)

  • demo a prebuilt version of the application

  • min-max normalization

  • 3 Nearest Neighbors euclidean distance calculation

  • code the application

  • go over training data used in this application

  • min-max normalization code

  • euclidean distance calculation code

  • Sentiment classification prediction code

  • handle command line entry code

  • test run completed application

  • use 3 nearest neighbors

  • normalize inputs and calculate euclidean distance

  • use 5 nearest neighbors

• Intro to Linear Algebra + Python NumPy (Prototype Project 01)

  • demo a prebuilt version of the application

  • add 2 matrices (both must have same size. add elements in same position)

  • subtract 2 matrices (both must have same size. subtract elements in same position)

  • multiply 2 matrices (both must have matching inner dimension. size of resulting matrix obtained by dropping middle dimension)

  • step by step example of multiplying 2 matrices

  • scalar add (operation applied to each element in matrix)

  • scalar subtract (operation applied to each element in matrix)

  • scalar multiply (operation applied to each element in matrix)

  • scalar divide (operation applied to each element in matrix)

  • identity matrix contains all 0s except diagonal is 1s (multiply matrix by identity matrix and obtain original matrix)

  • transpose matrix by flipping it along diagonal. rows become columns and columns become rows

  • dot product happens between 2 vectors (here we do element-wise multiplication than sum up results)

  • NumPy setup using miniconda

  • code the application

  • matrix as 2 dimensional array

  • add and subtract matrices in NumPy

  • multiply matrices in NumPy

  • matrix scalar operations in NumPy

  • identity matrix operation in NumPy

  • transpose matrix operation in NumPy

  • dot product matrix operation in NumPy

  • vector matrix