Abstract image of a woman sitting on books and studying on a laptop

M12 random forest-part01

Raman Kannan, profile picture
Raman Kannan

The document describes the process of analyzing a heart disease dataset using a random forest model. It includes details on feature selection, model training with different configurations, and performance metrics such as accuracy, sensitivity, and specificity. The results indicate various levels of predictive accuracy for different model setups, highlighting the effectiveness of the random forest approach in this context.

Data & Analytics
1 of 14
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
file<-'c://Users/rk215/Data/heart.csv' heart<-read.csv(file,head=T,sep=',',stringsAsFactors=F) unlist(lapply(names(heart),FUN=function(x,data=heart){c(cname=x,uvalfreq=len gth(unique(data[[x]])))})) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 1 of 14 11/23/2020, 5:44 PM
## cname uvalfreq cname uvalfreq cname uvalfreq cn ame ## "age" "41" "sex" "2" "cp" "4" "trestb ps" ## uvalfreq cname uvalfreq cname uvalfreq cname uvalf req ## "49" "chol" "152" "fbs" "2" "restecg" "3" ## cname uvalfreq cname uvalfreq cname uvalfreq cn ame ## "thalach" "91" "exang" "2" "oldpeak" "40" "slo pe" ## uvalfreq cname uvalfreq cname uvalfreq cname uvalf req ## "3" "ca" "5" "thal" "4" "target" "2" nvpairsdf3<-do.call('rbind',lapply(names(heart),FUN=function(x,data=heart){c (cname=x,uvalfreq=length(unique(data[[x]])))})) categoricalFeatures<-function(dset,ncol=7) { df4<-as.data.frame(do.call('rbind',lapply(names(dset),FUN=function(x,data= dset){c(cname=x,uvalfreq=length(unique(data[[x]])))}))) df4$uvalfreq=as.numeric(df4$uvalfreq) dset[,df4$uvalfreq<=ncol] } X<-categoricalFeatures(heart) table(heart$target) ## ## 0 1 ## 138 165 dim(X) ## [1] 303 9 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 2 of 14 11/23/2020, 5:44 PM
class_col<-which(names(X)=='target') fvcnt<-ncol(X[,-c(class_col)]) #set.seed if you want to repeatability #RF implementations consider sqrt(p) features # to avoid too many common features # here we are seeking to establish that it matters # we are concerned about features being present in both exp_fset1<-sample(1:fvcnt,fvcnt-1,replace=F) exp_fset2<-sample(1:fvcnt,fvcnt-1,replace=F) table(sort(exp_fset1)==sort(exp_fset2)) ## ## TRUE ## 7 exp_fset1 ## [1] 1 2 5 4 3 6 7 exp_fset2 ## [1] 7 3 1 2 5 6 4 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 3 of 14 11/23/2020, 5:44 PM
## ## rcpALL 0 1 ## 0 116 14 ## 1 22 151 ## [1] 0.8778656 ## [1] 0.0000000 0.1594203 1.0000000 ## [1] 0.0000000 0.9151515 1.0000000 X1<-X[,c(exp_fset1,9)] rpart.model1<-rpart(target~.,data=X1,minsplit=3) rpart.plot(rpart.model1) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 4 of 14 11/23/2020, 5:44 PM
rcp1<-predict(rpart.model1, X1[,-c(which(names(X1)=="target"))],type=" class") (rpart_mtab1<-table(rcp1,X$target)) ## ## rcp1 0 1 ## 0 107 15 ## 1 31 150 L1<-getMetrics(X1$target,as.numeric(rcp1)) L1$auc ## [1] 0.8422266 L1$fpr ## [1] 0.0000000 0.2246377 1.0000000 L1$tpr M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 5 of 14 11/23/2020, 5:44 PM
## [1] 0.0000000 0.9090909 1.0000000 X2<-X[,c(exp_fset2,9)] rpart.model2<-rpart(target~.,data=X2,minsplit=3) rpart.plot(rpart.model2) rcp2<-predict(rpart.model2, X2[,-c(which(names(X2)=="target"))],type=" class") (rpart_mtab2<-table(rcp2,X2$target)) ## ## rcp2 0 1 ## 0 107 15 ## 1 31 150 L2<-getMetrics(X2$target,as.numeric(rcp2)) L2$auc ## [1] 0.8422266 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 6 of 14 11/23/2020, 5:44 PM
L2$fpr ## [1] 0.0000000 0.2246377 1.0000000 L2$tpr ## [1] 0.0000000 0.9090909 1.0000000 exp_fset3<-sample(1:fvcnt,fvcnt-1,replace=F) X3<-X[,c(exp_fset3,9)] rpart.model3<-rpart(target~.,data=X3,minsplit=3) rpart.plot(rpart.model2) rcp3<-predict(rpart.model3, X3[,-c(which(names(X3)=="target"))],type=" class") (rpart_mtab3<-table(rcp3,X3$target)) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 7 of 14 11/23/2020, 5:44 PM
## ## rcp3 0 1 ## 0 111 24 ## 1 27 141 L3<-getMetrics(X3$target,as.numeric(rcp3)) L3$auc ## [1] 0.8294466 L3$fpr ## [1] 0.0000000 0.1956522 1.0000000 L3$tpr ## [1] 0.0000000 0.8545455 1.0000000 dfpred<-data.frame(actual=X$target,rcpALL,rcp1,rcp2,rcp3) head(dfpred) ## actual rcpALL rcp1 rcp2 rcp3 ## 1 1 1 1 1 1 ## 2 1 1 1 1 1 ## 3 1 1 1 1 1 ## 4 1 1 1 1 1 ## 5 1 1 1 1 0 ## 6 1 1 1 1 1 table(X$target,rcpALL) ## rcpALL ## 0 1 ## 0 116 22 ## 1 14 151 table(X$target,rcp1) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 8 of 14 11/23/2020, 5:44 PM
## rcp1 ## 0 1 ## 0 107 31 ## 1 15 150 table(X$target,rcp2) ## rcp2 ## 0 1 ## 0 107 31 ## 1 15 150 table(X$target,rcp3) ## rcp3 ## 0 1 ## 0 111 27 ## 1 24 141 … … if(!require(randomForest))require(randomForest) ## Loading required package: randomForest ## randomForest 4.6-14 ## Type rfNews() to see new features/changes/bug fixes. rf_model<-randomForest(target~.,data=X) rf_pred<-predict(rf_model,X[,-c(which(names(X)=="target"))]) rf_mtab<-table(X$target,rf_pred) rf_cmx<-caret::confusionMatrix(rf_mtab) rf_mtab ## rf_pred ## 0 1 ## 0 124 14 ## 1 10 155 (rf_accuracy<-sum(diag(rf_mtab))/sum(rf_mtab)) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 9 of 14 11/23/2020, 5:44 PM
## [1] 0.9207921 rf_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.207921e-01 8.399366e-01 8.844364e-01 9.485923e-01 5.577558e-0 1 ## AccuracyPValue McnemarPValue ## 1.441683e-44 5.402914e-01 rf_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9253731 0.9171598 0.8985507 ## Neg Pred Value Precision Recall ## 0.9393939 0.8985507 0.9253731 ## F1 Prevalence Detection Rate ## 0.9117647 0.4422442 0.4092409 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9212664 rf1k_model<-randomForest(target~.,data=X,ntree=1000) rf1k_pred<-predict(rf1k_model,X[,-c(which(names(X)=="target"))]) rf1k_mtab<-table(X$target,rf1k_pred) rf1k_cmx<-caret::confusionMatrix(rf1k_mtab) rf1k_mtab ## rf1k_pred ## 0 1 ## 0 121 17 ## 1 8 157 (rf1k_accuracy<-sum(diag(rf1k_mtab))/sum(rf1k_mtab)) ## [1] 0.9174917 rf1k_cmx$overall M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 10 of 14 11/23/2020, 5:44 PM
## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.174917e-01 8.327704e-01 8.806106e-01 9.458942e-01 5.742574e-0 1 ## AccuracyPValue McnemarPValue ## 1.656736e-40 1.095986e-01 rf1k_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9379845 0.9022989 0.8768116 ## Neg Pred Value Precision Recall ## 0.9515152 0.8768116 0.9379845 ## F1 Prevalence Detection Rate ## 0.9063670 0.4257426 0.3993399 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9201417 rf100_model<-randomForest(target~.,data=X,ntree=100) rf100_pred<-predict(rf100_model,X[,-c(which(names(X)=="target"))]) rf100_mtab<-table(X$target,rf100_pred) rf100_cmx<-caret::confusionMatrix(rf100_mtab) rf100_mtab ## rf100_pred ## 0 1 ## 0 123 15 ## 1 11 154 (rf100_accuracy<-sum(diag(rf100_mtab))/sum(rf100_mtab)) ## [1] 0.9141914 rf100_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.141914e-01 8.265980e-01 8.767989e-01 9.431813e-01 5.577558e-0 1 ## AccuracyPValue McnemarPValue ## 1.093523e-42 5.562985e-01 rf100_cmx$byClass M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 11 of 14 11/23/2020, 5:44 PM
## Sensitivity Specificity Pos Pred Value ## 0.9179104 0.9112426 0.8913043 ## Neg Pred Value Precision Recall ## 0.9333333 0.8913043 0.9179104 ## F1 Prevalence Detection Rate ## 0.9044118 0.4422442 0.4059406 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9145765 rf50_model<-randomForest(target~.,data=X,ntree=50) rf50_pred<-predict(rf50_model,X[,-c(which(names(X)=="target"))]) rf50_mtab<-table(X$target,rf50_pred) rf50_cmx<-caret::confusionMatrix(rf50_mtab) rf50_mtab ## rf50_pred ## 0 1 ## 0 124 14 ## 1 11 154 (rf1k_accuracy<-sum(diag(rf50_mtab))/sum(rf50_mtab)) ## [1] 0.9174917 rf1k_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.174917e-01 8.327704e-01 8.806106e-01 9.458942e-01 5.742574e-0 1 ## AccuracyPValue McnemarPValue ## 1.656736e-40 1.095986e-01 rf1k_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9379845 0.9022989 0.8768116 ## Neg Pred Value Precision Recall ## 0.9515152 0.8768116 0.9379845 ## F1 Prevalence Detection Rate ## 0.9063670 0.4257426 0.3993399 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9201417 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 12 of 14 11/23/2020, 5:44 PM
rfgrow<-grow(rf50_model,50) rfgrow_pred<-predict(rfgrow,X[,-c(which(names(X)=="target"))]) rfgrow_mtab<-table(X$target,rfgrow_pred) rfgrow_cmx<-caret::confusionMatrix(rfgrow_mtab) rfgrow_mtab ## rfgrow_pred ## 0 1 ## 0 123 15 ## 1 9 156 (rfgrow_accuracy<-sum(diag(rfgrow_mtab))/sum(rfgrow_mtab)) ## [1] 0.9207921 rfgrow_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.207921e-01 8.397461e-01 8.844364e-01 9.485923e-01 5.643564e-0 1 ## AccuracyPValue McnemarPValue ## 2.685766e-43 3.074342e-01 rfgrow_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9318182 0.9122807 0.8913043 ## Neg Pred Value Precision Recall ## 0.9454545 0.8913043 0.9318182 ## F1 Prevalence Detection Rate ## 0.9111111 0.4356436 0.4059406 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9220494 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 13 of 14 11/23/2020, 5:44 PM
M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 14 of 14 11/23/2020, 5:44 PM

More Related Content

PDF
M11 bagging loo cv
Raman Kannan
 
PDF
M09-Cross validating-naive-bayes
Raman Kannan
 
PDF
Chapter 2: R tutorial Handbook for Data Science and Machine Learning Practiti...
Raman Kannan
 
KEY
実践 memcached
Masahiro Nagano
 
PDF
Comparative Genomics with GMOD and BioPerl
Jason Stajich
 
ODP
Caching and tuning fun for high scalability @ LOAD2012
Wim Godden
 
PDF
Visualization of Supervised Learning with {arules} + {arulesViz}
Takashi J OZAKI
 
PDF
Redis 101
Doğan Can
 
M11 bagging loo cv
Raman Kannan
 
M09-Cross validating-naive-bayes
Raman Kannan
 
Chapter 2: R tutorial Handbook for Data Science and Machine Learning Practiti...
Raman Kannan
 
実践 memcached
Masahiro Nagano
 
Comparative Genomics with GMOD and BioPerl
Jason Stajich
 
Caching and tuning fun for high scalability @ LOAD2012
Wim Godden
 
Visualization of Supervised Learning with {arules} + {arulesViz}
Takashi J OZAKI
 
Redis 101
Doğan Can
 

What's hot (19)

TXT
Command
gajshield
 
PPTX
Five
Łukasz Langa
 
PDF
Association Rule Mining with R
Yanchang Zhao
 
ODP
Beyond php - it's not (just) about the code
Wim Godden
 
PPTX
Data manipulation and visualization in r 20190711 myanmarucsy
SmartHinJ
 
PPTX
Gevent rabbit rpc
Aleksandr Mokrov
 
ODP
Beyond php it's not (just) about the code
Wim Godden
 
PDF
Regression and Classification with R
Yanchang Zhao
 
PDF
Bytes in the Machine: Inside the CPython interpreter
akaptur
 
PDF
Allison Kaptur: Bytes in the Machine: Inside the CPython interpreter, PyGotha...
akaptur
 
PDF
Parallel Computing With Dask - PyDays 2017
Christian Aichinger
 
PPTX
System Calls
David Evans
 
PDF
Cooking pies with Celery
Aleksandr Mokrov
 
PDF
Sangam 19 - Successful Applications on Autonomous
Connor McDonald
 
PDF
Sangam 19 - PLSQL still the coolest
Connor McDonald
 
ODP
Beyond PHP - It's not (just) about the code
Wim Godden
 
PDF
Python testing-frameworks overview
Jachym Cepicky
 
PDF
python高级内存管理
rfyiamcool
 
PDF
Agile Database Development with JSON
Chris Saxon
 
Command
gajshield
 
Five
Łukasz Langa
 
Association Rule Mining with R
Yanchang Zhao
 
Beyond php - it's not (just) about the code
Wim Godden
 
Data manipulation and visualization in r 20190711 myanmarucsy
SmartHinJ
 
Gevent rabbit rpc
Aleksandr Mokrov
 
Beyond php it's not (just) about the code
Wim Godden
 
Regression and Classification with R
Yanchang Zhao
 
Bytes in the Machine: Inside the CPython interpreter
akaptur
 
Allison Kaptur: Bytes in the Machine: Inside the CPython interpreter, PyGotha...
akaptur
 
Parallel Computing With Dask - PyDays 2017
Christian Aichinger
 
System Calls
David Evans
 
Cooking pies with Celery
Aleksandr Mokrov
 
Sangam 19 - Successful Applications on Autonomous
Connor McDonald
 
Sangam 19 - PLSQL still the coolest
Connor McDonald
 
Beyond PHP - It's not (just) about the code
Wim Godden
 
Python testing-frameworks overview
Jachym Cepicky
 
python高级内存管理
rfyiamcool
 
Agile Database Development with JSON
Chris Saxon
 
Ad

Similar to M12 random forest-part01 (20)

DOCX
R (Shiny Package) - Server Side Code for Decision Support System
Maithreya Chakravarthula
 
PDF
Spark + Clojure for Topic Discovery - Zalando Tech Clojure/Conj Talk
Zalando Technology
 
PDF
R is a very flexible and powerful programming language, as well as a.pdf
annikasarees
 
PPTX
Introduction to R
Sander Kieft
 
PDF
NCCU: Statistics in the Criminal Justice System, R basics and Simulation - Pr...
The Statistical and Applied Mathematical Sciences Institute
 
PDF
Being functional in PHP (PHPDay Italy 2016)
David de Boer
 
PPTX
Seminar PSU 10.10.2014 mme
Vyacheslav Arbuzov
 
PDF
R code for data manipulation
Avjinder (Avi) Kaler
 
PDF
R code for data manipulation
Avjinder (Avi) Kaler
 
PDF
Solutions for Exercises in Digital Signal Processing Using MATLAB, 4th Editio...
electricaleng2024
 
DOCX
A Shiny Example-- R
Dr. Volkan OBAN
 
PPT
Vcs16
Malikireddy Bramhananda Reddy
 
DOCX
Ns2programs
Meenakshi Devi
 
PDF
Do snow.rwn
ARUN DN
 
TXT
R code
Manav Goel
 
PDF
Rcommands-for those who interested in R.
Dr. Volkan OBAN
 
PDF
Pre-Bootcamp introduction to Elixir
Paweł Dawczak
 
PPTX
Php functions
JIGAR MAKHIJA
 
PDF
Prediction
Ban Bang
 
PDF
Time Series Analysis and Mining with R
Yanchang Zhao
 
R (Shiny Package) - Server Side Code for Decision Support System
Maithreya Chakravarthula
 
Spark + Clojure for Topic Discovery - Zalando Tech Clojure/Conj Talk
Zalando Technology
 
R is a very flexible and powerful programming language, as well as a.pdf
annikasarees
 
Introduction to R
Sander Kieft
 
NCCU: Statistics in the Criminal Justice System, R basics and Simulation - Pr...
The Statistical and Applied Mathematical Sciences Institute
 
Being functional in PHP (PHPDay Italy 2016)
David de Boer
 
Seminar PSU 10.10.2014 mme
Vyacheslav Arbuzov
 
R code for data manipulation
Avjinder (Avi) Kaler
 
R code for data manipulation
Avjinder (Avi) Kaler
 
Solutions for Exercises in Digital Signal Processing Using MATLAB, 4th Editio...
electricaleng2024
 
A Shiny Example-- R
Dr. Volkan OBAN
 
Vcs16
Malikireddy Bramhananda Reddy
 
Ns2programs
Meenakshi Devi
 
Do snow.rwn
ARUN DN
 
R code
Manav Goel
 
Rcommands-for those who interested in R.
Dr. Volkan OBAN
 
Pre-Bootcamp introduction to Elixir
Paweł Dawczak
 
Php functions
JIGAR MAKHIJA
 
Prediction
Ban Bang
 
Time Series Analysis and Mining with R
Yanchang Zhao
 
Ad

More from Raman Kannan (20)

PDF
conversations-withchatGPT-Claude-gemini-on-vibration-12-112024.pdf
Raman Kannan
 
PDF
Essays on-civic-responsibilty
Raman Kannan
 
PDF
M12 boosting-part02
Raman Kannan
 
PDF
M10 gradient descent
Raman Kannan
 
PDF
M06 tree
Raman Kannan
 
PDF
M07 svm
Raman Kannan
 
PDF
M08 BiasVarianceTradeoff
Raman Kannan
 
PDF
Chapter 05 k nn
Raman Kannan
 
PDF
Chapter 04-discriminant analysis
Raman Kannan
 
PDF
M03 nb-02
Raman Kannan
 
PDF
Augmented 11022020-ieee
Raman Kannan
 
PDF
Chapter 02-logistic regression
Raman Kannan
 
PDF
Chapter01 introductory handbook
Raman Kannan
 
PDF
A voyage-inward-02
Raman Kannan
 
PDF
Evaluating classifierperformance ml-cs6923
Raman Kannan
 
PDF
A data scientist's study plan
Raman Kannan
 
PDF
Cognitive Assistants
Raman Kannan
 
PDF
Essay on-data-analysis
Raman Kannan
 
PDF
Joy of Unix
Raman Kannan
 
PDF
How to-run-ols-diagnostics-02
Raman Kannan
 
conversations-withchatGPT-Claude-gemini-on-vibration-12-112024.pdf
Raman Kannan
 
Essays on-civic-responsibilty
Raman Kannan
 
M12 boosting-part02
Raman Kannan
 
M10 gradient descent
Raman Kannan
 
M06 tree
Raman Kannan
 
M07 svm
Raman Kannan
 
M08 BiasVarianceTradeoff
Raman Kannan
 
Chapter 05 k nn
Raman Kannan
 
Chapter 04-discriminant analysis
Raman Kannan
 
M03 nb-02
Raman Kannan
 
Augmented 11022020-ieee
Raman Kannan
 
Chapter 02-logistic regression
Raman Kannan
 
Chapter01 introductory handbook
Raman Kannan
 
A voyage-inward-02
Raman Kannan
 
Evaluating classifierperformance ml-cs6923
Raman Kannan
 
A data scientist's study plan
Raman Kannan
 
Cognitive Assistants
Raman Kannan
 
Essay on-data-analysis
Raman Kannan
 
Joy of Unix
Raman Kannan
 
How to-run-ols-diagnostics-02
Raman Kannan
 

Recently uploaded (20)

PPTX
CYBER SECURITY the Next Warefare Tactics
AbdullahAaliyan
 
PPTX
(Ali Hamza) Roll No: (F24-BSCS-1103).pptx
fb7654821
 
PDF
Data Engineering Interview Questions & Answers Batch Processing (Spark, Hadoo...
Mindbox Trainings
 
PPTX
New ISO 27001_2022 standard and the changes
Mayank Mathur
 
PDF
Introduction to the R Programming Language
Suraj Patil
 
PDF
Systems Analysis and Design, 12th Edition by Scott Tilley Test Bank.pdf
solutionsmanual3
 
PDF
Navigating the Thai Supplements Landscape.pdf
TumNana
 
PDF
REAL ILLUMINATI AGENT IN KAMPALA UGANDA CALL ON+256765750853/0705037305
Real illuminati agent from Uganda Kampala call 0782561496/0756664682
 
PPT
lectureusjsjdhdsjjshdshshddhdhddhhd1.ppt
dipa69
 
PPTX
Managing Community Partner Relationships
alexmderr
 
PPT
DU, AIS, Big Data and Data Analytics.ppt
SadihaAfrose
 
PPTX
Business_Capability_Map_Collection__pptx
anandazad4
 
PPT
Image processing and pattern recognition 2.ppt
anwarfadel430
 
PPTX
Pilar Kemerdekaan dan Identi Bangsa.pptx
RyanSyah40
 
PPTX
Lesson-01intheselfoflifeofthekennyrogersoftheunderstandoftheunderstanded
adipaul20
 
PPTX
Copy of 16 Timeline & Flowchart Templates – HubSpot.pptx
ahsonbashir
 
PPTX
Introduction to Inferential Statistics.pptx
CollegeofTeacherEduc3
 
PDF
Votre score augmente si vous choisissez une catégorie et que vous rédigez une...
GreatIdeasin5Minutes
 
PDF
Tetra Pak Index 2023 - The future of health and nutrition - Full report.pdf
PhmK5
 
PPTX
DS-40-Pre-Engagement and Kickoff deck - v8.0.pptx
GuillermoIgnacioAlva
 
CYBER SECURITY the Next Warefare Tactics
AbdullahAaliyan
 
(Ali Hamza) Roll No: (F24-BSCS-1103).pptx
fb7654821
 
Data Engineering Interview Questions & Answers Batch Processing (Spark, Hadoo...
Mindbox Trainings
 
New ISO 27001_2022 standard and the changes
Mayank Mathur
 
Introduction to the R Programming Language
Suraj Patil
 
Systems Analysis and Design, 12th Edition by Scott Tilley Test Bank.pdf
solutionsmanual3
 
Navigating the Thai Supplements Landscape.pdf
TumNana
 
REAL ILLUMINATI AGENT IN KAMPALA UGANDA CALL ON+256765750853/0705037305
Real illuminati agent from Uganda Kampala call 0782561496/0756664682
 
lectureusjsjdhdsjjshdshshddhdhddhhd1.ppt
dipa69
 
Managing Community Partner Relationships
alexmderr
 
DU, AIS, Big Data and Data Analytics.ppt
SadihaAfrose
 
Business_Capability_Map_Collection__pptx
anandazad4
 
Image processing and pattern recognition 2.ppt
anwarfadel430
 
Pilar Kemerdekaan dan Identi Bangsa.pptx
RyanSyah40
 
Lesson-01intheselfoflifeofthekennyrogersoftheunderstandoftheunderstanded
adipaul20
 
Copy of 16 Timeline & Flowchart Templates – HubSpot.pptx
ahsonbashir
 
Introduction to Inferential Statistics.pptx
CollegeofTeacherEduc3
 
Votre score augmente si vous choisissez une catégorie et que vous rédigez une...
GreatIdeasin5Minutes
 
Tetra Pak Index 2023 - The future of health and nutrition - Full report.pdf
PhmK5
 
DS-40-Pre-Engagement and Kickoff deck - v8.0.pptx
GuillermoIgnacioAlva
 

M12 random forest-part01

  • 2. ## cname uvalfreq cname uvalfreq cname uvalfreq cn ame ## "age" "41" "sex" "2" "cp" "4" "trestb ps" ## uvalfreq cname uvalfreq cname uvalfreq cname uvalf req ## "49" "chol" "152" "fbs" "2" "restecg" "3" ## cname uvalfreq cname uvalfreq cname uvalfreq cn ame ## "thalach" "91" "exang" "2" "oldpeak" "40" "slo pe" ## uvalfreq cname uvalfreq cname uvalfreq cname uvalf req ## "3" "ca" "5" "thal" "4" "target" "2" nvpairsdf3<-do.call('rbind',lapply(names(heart),FUN=function(x,data=heart){c (cname=x,uvalfreq=length(unique(data[[x]])))})) categoricalFeatures<-function(dset,ncol=7) { df4<-as.data.frame(do.call('rbind',lapply(names(dset),FUN=function(x,data= dset){c(cname=x,uvalfreq=length(unique(data[[x]])))}))) df4$uvalfreq=as.numeric(df4$uvalfreq) dset[,df4$uvalfreq<=ncol] } X<-categoricalFeatures(heart) table(heart$target) ## ## 0 1 ## 138 165 dim(X) ## [1] 303 9 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 2 of 14 11/23/2020, 5:44 PM
  • 3. class_col<-which(names(X)=='target') fvcnt<-ncol(X[,-c(class_col)]) #set.seed if you want to repeatability #RF implementations consider sqrt(p) features # to avoid too many common features # here we are seeking to establish that it matters # we are concerned about features being present in both exp_fset1<-sample(1:fvcnt,fvcnt-1,replace=F) exp_fset2<-sample(1:fvcnt,fvcnt-1,replace=F) table(sort(exp_fset1)==sort(exp_fset2)) ## ## TRUE ## 7 exp_fset1 ## [1] 1 2 5 4 3 6 7 exp_fset2 ## [1] 7 3 1 2 5 6 4 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 3 of 14 11/23/2020, 5:44 PM
  • 4. ## ## rcpALL 0 1 ## 0 116 14 ## 1 22 151 ## [1] 0.8778656 ## [1] 0.0000000 0.1594203 1.0000000 ## [1] 0.0000000 0.9151515 1.0000000 X1<-X[,c(exp_fset1,9)] rpart.model1<-rpart(target~.,data=X1,minsplit=3) rpart.plot(rpart.model1) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 4 of 14 11/23/2020, 5:44 PM
  • 5. rcp1<-predict(rpart.model1, X1[,-c(which(names(X1)=="target"))],type=" class") (rpart_mtab1<-table(rcp1,X$target)) ## ## rcp1 0 1 ## 0 107 15 ## 1 31 150 L1<-getMetrics(X1$target,as.numeric(rcp1)) L1$auc ## [1] 0.8422266 L1$fpr ## [1] 0.0000000 0.2246377 1.0000000 L1$tpr M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 5 of 14 11/23/2020, 5:44 PM
  • 6. ## [1] 0.0000000 0.9090909 1.0000000 X2<-X[,c(exp_fset2,9)] rpart.model2<-rpart(target~.,data=X2,minsplit=3) rpart.plot(rpart.model2) rcp2<-predict(rpart.model2, X2[,-c(which(names(X2)=="target"))],type=" class") (rpart_mtab2<-table(rcp2,X2$target)) ## ## rcp2 0 1 ## 0 107 15 ## 1 31 150 L2<-getMetrics(X2$target,as.numeric(rcp2)) L2$auc ## [1] 0.8422266 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 6 of 14 11/23/2020, 5:44 PM
  • 7. L2$fpr ## [1] 0.0000000 0.2246377 1.0000000 L2$tpr ## [1] 0.0000000 0.9090909 1.0000000 exp_fset3<-sample(1:fvcnt,fvcnt-1,replace=F) X3<-X[,c(exp_fset3,9)] rpart.model3<-rpart(target~.,data=X3,minsplit=3) rpart.plot(rpart.model2) rcp3<-predict(rpart.model3, X3[,-c(which(names(X3)=="target"))],type=" class") (rpart_mtab3<-table(rcp3,X3$target)) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 7 of 14 11/23/2020, 5:44 PM
  • 8. ## ## rcp3 0 1 ## 0 111 24 ## 1 27 141 L3<-getMetrics(X3$target,as.numeric(rcp3)) L3$auc ## [1] 0.8294466 L3$fpr ## [1] 0.0000000 0.1956522 1.0000000 L3$tpr ## [1] 0.0000000 0.8545455 1.0000000 dfpred<-data.frame(actual=X$target,rcpALL,rcp1,rcp2,rcp3) head(dfpred) ## actual rcpALL rcp1 rcp2 rcp3 ## 1 1 1 1 1 1 ## 2 1 1 1 1 1 ## 3 1 1 1 1 1 ## 4 1 1 1 1 1 ## 5 1 1 1 1 0 ## 6 1 1 1 1 1 table(X$target,rcpALL) ## rcpALL ## 0 1 ## 0 116 22 ## 1 14 151 table(X$target,rcp1) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 8 of 14 11/23/2020, 5:44 PM
  • 9. ## rcp1 ## 0 1 ## 0 107 31 ## 1 15 150 table(X$target,rcp2) ## rcp2 ## 0 1 ## 0 107 31 ## 1 15 150 table(X$target,rcp3) ## rcp3 ## 0 1 ## 0 111 27 ## 1 24 141 … … if(!require(randomForest))require(randomForest) ## Loading required package: randomForest ## randomForest 4.6-14 ## Type rfNews() to see new features/changes/bug fixes. rf_model<-randomForest(target~.,data=X) rf_pred<-predict(rf_model,X[,-c(which(names(X)=="target"))]) rf_mtab<-table(X$target,rf_pred) rf_cmx<-caret::confusionMatrix(rf_mtab) rf_mtab ## rf_pred ## 0 1 ## 0 124 14 ## 1 10 155 (rf_accuracy<-sum(diag(rf_mtab))/sum(rf_mtab)) M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 9 of 14 11/23/2020, 5:44 PM
  • 10. ## [1] 0.9207921 rf_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.207921e-01 8.399366e-01 8.844364e-01 9.485923e-01 5.577558e-0 1 ## AccuracyPValue McnemarPValue ## 1.441683e-44 5.402914e-01 rf_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9253731 0.9171598 0.8985507 ## Neg Pred Value Precision Recall ## 0.9393939 0.8985507 0.9253731 ## F1 Prevalence Detection Rate ## 0.9117647 0.4422442 0.4092409 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9212664 rf1k_model<-randomForest(target~.,data=X,ntree=1000) rf1k_pred<-predict(rf1k_model,X[,-c(which(names(X)=="target"))]) rf1k_mtab<-table(X$target,rf1k_pred) rf1k_cmx<-caret::confusionMatrix(rf1k_mtab) rf1k_mtab ## rf1k_pred ## 0 1 ## 0 121 17 ## 1 8 157 (rf1k_accuracy<-sum(diag(rf1k_mtab))/sum(rf1k_mtab)) ## [1] 0.9174917 rf1k_cmx$overall M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 10 of 14 11/23/2020, 5:44 PM
  • 11. ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.174917e-01 8.327704e-01 8.806106e-01 9.458942e-01 5.742574e-0 1 ## AccuracyPValue McnemarPValue ## 1.656736e-40 1.095986e-01 rf1k_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9379845 0.9022989 0.8768116 ## Neg Pred Value Precision Recall ## 0.9515152 0.8768116 0.9379845 ## F1 Prevalence Detection Rate ## 0.9063670 0.4257426 0.3993399 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9201417 rf100_model<-randomForest(target~.,data=X,ntree=100) rf100_pred<-predict(rf100_model,X[,-c(which(names(X)=="target"))]) rf100_mtab<-table(X$target,rf100_pred) rf100_cmx<-caret::confusionMatrix(rf100_mtab) rf100_mtab ## rf100_pred ## 0 1 ## 0 123 15 ## 1 11 154 (rf100_accuracy<-sum(diag(rf100_mtab))/sum(rf100_mtab)) ## [1] 0.9141914 rf100_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.141914e-01 8.265980e-01 8.767989e-01 9.431813e-01 5.577558e-0 1 ## AccuracyPValue McnemarPValue ## 1.093523e-42 5.562985e-01 rf100_cmx$byClass M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 11 of 14 11/23/2020, 5:44 PM
  • 12. ## Sensitivity Specificity Pos Pred Value ## 0.9179104 0.9112426 0.8913043 ## Neg Pred Value Precision Recall ## 0.9333333 0.8913043 0.9179104 ## F1 Prevalence Detection Rate ## 0.9044118 0.4422442 0.4059406 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9145765 rf50_model<-randomForest(target~.,data=X,ntree=50) rf50_pred<-predict(rf50_model,X[,-c(which(names(X)=="target"))]) rf50_mtab<-table(X$target,rf50_pred) rf50_cmx<-caret::confusionMatrix(rf50_mtab) rf50_mtab ## rf50_pred ## 0 1 ## 0 124 14 ## 1 11 154 (rf1k_accuracy<-sum(diag(rf50_mtab))/sum(rf50_mtab)) ## [1] 0.9174917 rf1k_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.174917e-01 8.327704e-01 8.806106e-01 9.458942e-01 5.742574e-0 1 ## AccuracyPValue McnemarPValue ## 1.656736e-40 1.095986e-01 rf1k_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9379845 0.9022989 0.8768116 ## Neg Pred Value Precision Recall ## 0.9515152 0.8768116 0.9379845 ## F1 Prevalence Detection Rate ## 0.9063670 0.4257426 0.3993399 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9201417 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 12 of 14 11/23/2020, 5:44 PM
  • 13. rfgrow<-grow(rf50_model,50) rfgrow_pred<-predict(rfgrow,X[,-c(which(names(X)=="target"))]) rfgrow_mtab<-table(X$target,rfgrow_pred) rfgrow_cmx<-caret::confusionMatrix(rfgrow_mtab) rfgrow_mtab ## rfgrow_pred ## 0 1 ## 0 123 15 ## 1 9 156 (rfgrow_accuracy<-sum(diag(rfgrow_mtab))/sum(rfgrow_mtab)) ## [1] 0.9207921 rfgrow_cmx$overall ## Accuracy Kappa AccuracyLower AccuracyUpper AccuracyNul l ## 9.207921e-01 8.397461e-01 8.844364e-01 9.485923e-01 5.643564e-0 1 ## AccuracyPValue McnemarPValue ## 2.685766e-43 3.074342e-01 rfgrow_cmx$byClass ## Sensitivity Specificity Pos Pred Value ## 0.9318182 0.9122807 0.8913043 ## Neg Pred Value Precision Recall ## 0.9454545 0.8913043 0.9318182 ## F1 Prevalence Detection Rate ## 0.9111111 0.4356436 0.4059406 ## Detection Prevalence Balanced Accuracy ## 0.4554455 0.9220494 M12-RandomForest file:///E:/users/rkannan/cuny/fall2020/fall2020/ML-Handbook/m12-rFore... 13 of 14 11/23/2020, 5:44 PM