Homework 1

Homework 1 August 31 st , 2022 # QUESTION 2.1 Recently I was talking with the Real-Estate department team from the company that I’m working at. It is very interesting how SVM and classifications could play a huge role. When a company trying to open a new branch, or invest in a location or a building, there are a few predictors to be determined. Such as, the median income of that location, market price, force majeure risks, number of populations, and crime rates. # QUESTION 2.2 library ( kernlab ) library ( rsample ) library ( caret ) creditdata = read.table ( "credit_card_data.txt" , header = FALSE , stringsAsFactors = FALSE ) #import data data = as.matrix ( creditdata ) head ( creditdata ) #to sample the first small sample of rows present in the input data frame tail ( creditdata ) #to sample the last small sample of rows present in the input data frame svm_c = function ( c ){ #defining a function model = ksvm ( data [ , 1 : 10 ] ,as.factor ( data [ , 11 ]) ,type = "C- svc" , kernel = "vanilladot" ,C = lambda,scaled = TRUE ) #call ksvm pred = predict ( model, data [ , 1 : 10 ]) #see what the model predicts sum ( pred == data [ , 11 ]) / nrow ( data ) #see what fraction of the model’s predictions match the actual classification } print ( svm_c ( 10 )) #Checking for the accuracy of the model print ( svm_c ( 100 )) #Checking for the accuracy of the model print ( svm_c ( 1000 )) #hecking for the accuracy of the model # Using 100 as the best C model = ksvm ( data [ , 1 : 10 ] ,as.factor ( data [ , 11 ]) ,type = "C- svc" , kernel = "vanilladot" ,C = 100 ,scaled = TRUE ) calculate_a = colSums ( model@xmatrix [[ 1 ]] * model@ coef [[ 1 ]]) #calculate a1 ... am print ( calculate_a ) calculate_a0 = model@b #calculate a0 print ( calculate_a0 ) # -0.08158492 pred = predict ( model, data [ , 1 : 10 ]) #getting the prediction print ( sum ( pred == data [ , 11 ]) / nrow ( data )) #86.39%