upload

upload files

.cache/__packages

@@ -0,0 +1,29 @@
+base
+BBmisc
+devtools
+stringr
+stringi
+reshape
+reshape2
+data.table
+plyr
+dplyr
+magrittr
+foreach
+iterators
+doParallel
+knitr
+rmarkdown
+tidyr
+gtable
+gridExtra
+pander
+stringdist
+slidify
+RColorBrewer
+leaflet
+installr
+plot3D
+markdown
+broman
+foreign

.gitignore

@@ -0,0 +1,14 @@
+# History files
+.Rhistory
+.Rapp.history
+
+# Example code in package build process
+*-Ex.R
+
+# RStudio files
+.Rproj.user/
+
+# produced vignettes
+vignettes/*.html
+vignettes/*.pdf
+.Rproj.user

Betting-Strategy-and-Model-Validation.Rproj

@@ -0,0 +1,13 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX

Natural_Language_Analysis_cache/html/__packages

@@ -0,0 +1,31 @@
+base
+methods
+datasets
+utils
+grDevices
+graphics
+stats
+scimapClient
+BBmisc
+devtools
+stringr
+stringi
+reshape
+reshape2
+data.table
+DT
+plyr
+dplyr
+magrittr
+foreach
+iterators
+parallel
+doParallel
+rmarkdown
+tidyr
+grid
+gtable
+gridExtra
+pander
+stringdist
+knitr

PL.R

@@ -0,0 +1,170 @@
+## Loading the packages
+if(!'BBmisc' %in% installed.packages()){
+  install.packages('BBmisc')}
+if(!'BiocParallel' %in% installed.packages()){
+  source("http://bioconductor.org/biocLite.R")
+  biocLite("BiocParallel")}
+if(!'seleniumJars' %in% installed.packages()){
+  install_github('LluisRamon/seleniumJars')}
+
+suppressPackageStartupMessages(library('BBmisc'))
+suppressPackageStartupMessages(lib(c('zoo','stringi','stringr','reshape','reshape2','plyr','dplyr','magrittr',
+                                     'ggplot2','ggthemes','plotly','foreach','memoise','doMC','doParallel','BiocParallel',
+                                     'markdown','parallel','rmarkdown','manipulate','knitr','turner','scales',
+                                     'lubridate','whisker'))) #'RStudioAMI','editR'
+
+## ---------------------------------------------------------------------------------------------
+## http://stackoverflow.com/questions/22954623/view-markdown-generated-html-in-rstudio-viewer
+render(paste0(getwd(),'/Betting Strategy and Model Validation.Rmd'),'all')
+#'@ View(paste0(getwd(),'/Betting_Strategy_and_Model_Validation.html'))
+browseURL(paste0(getwd(),'/Betting Strategy and Model Validation.html'))
+
+## https://github.com/swarm-lab/editR
+editR(paste0(getwd(),'/Betting Strategy and Model Validation.Rmd'))
+
+
+## Besides, need to scrap the final-scores / half-time scores / result of soccer matches
+teamID <- sort(unique(c(as.character(mbase$Home), as.character(mbase$Away))))
+dateID <- sort(unique(mbase$Date)); spboDate <- gsub('-','',dateID)
+lnk <- paste0('http://www8.spbo.com/history.plex?day=',spboDate,'&l=en')
+
+## http://stackoverflow.com/questions/2158780/r-catching-an-error-and-then-branching-logic
+## http://www.win-vector.com/blog/2012/10/error-handling-in-r/
+## Due to the scrapSPBO function scrapped unmatched data, example lnk[827],
+##  therefore I rewrite the function as scrapSPBO2
+source(paste0(getwd(),'/function/scrapSPBO2.R'))
+scrapSPBO2(lnk=lnk, dateID=dateID, path='livescore', parallel=FALSE)
+
+## Read scraped spbo datasets
+source(paste0(getwd(),'/function/readSPBO.R'))
+spboData <- readSPBO(dateID=dateID, path='livescore', parallel=FALSE)
+
+
+
+## https://github.com/pablobarbera/instaR
+## https://github.com/pablobarbera/Rfacebook
+install_github ## can try during free time
+## ---------------------------------------------------------------------------------------------
+## Load the scraped spbo livescore datasets.
+##... will take some times since dim spboData [156841 x 17]
+source(paste0(getwd(),'/function/readSPBO2.R'))
+spboData <- readSPBO2(dateID=dateID, parallel=TRUE)
+
+## filter spboTeamID
+spboTeamID <- sort(c(unique(as.vector(spboData$Home)),unique(as.vector(spboData$Away))))
+tmID <- teamID[!teamID %in% mbase$others]
+
+spboData[(is.na(spboData$Date))&(nchar(as.vector(spboData$Time))==5),]
+spboData[subset(spboData, (is.na(data.frame(spboData)$Date))&(nchar(as.vector(spboData$Time))==5))$X,]
+
+> dim(mbase$datasets)
+[1] 48744    17
+> dim(spboData)
+[1] 319744     20
+
+mbase$datasets[mbase$datasets$DateUK %in% spboData$DateUK,]
+#Source: local data frame [17,934 x 17]
+
+na.omit(mbase$datasets[mbase$datasets$DateUK %in% spboData$DateUK,][order(mbase$datasets$No,decreasing=FALSE),])
+#Source: local data frame [25,489 x 17]
+
+library('tau')
+library('textcat')
+library('stringdist')
+
+
+
+
+
+
+
+
+
+
+
+http://wizardofvegas.com/forum/gambling/sports/10555-halt-time-betting/3/
+http://quant.stackexchange.com/questions/2500/how-to-apply-the-kelly-criterion-when-expected-return-may-be-negative
+https://en.wikipedia.org/wiki/Gambling_and_information_theory
+http://www.eecs.harvard.edu/cs286r/courses/fall12/papers/Thorpe_KellyCriterion2007.pdf
+http://www.sportsbookreview.com/betting-tools/kelly-calculator/
+http://thestakingmachine.com/laykelly.php
+### http://www.sportsbettingcalculator.co.uk/kelly-staking-calculator/
+http://tipstertables.com/blog/betting-system-using-tipster-statistics-and-kelly-criterion
+########################################################################################
+
+## Scrape the League in order to assign the virogish/spread margins/overrounds
+library(RSelenium)
+teamID <- sort(unique(unlist(mbase$Home), unlist(mbase$Away)))
+lnk <- 'http://www8.spbo.com/history.plex?day=20110107&l=en'
+
+#'@ system('java -jar selenium-server-standalone.jar')
+checkForServer() ## if you need the stand-alone Java binary
+startServer()
+webDr <- remoteDriver$new()
+webDr$open()
+webDr$navigate(lnk)
+webDr$navigate("http://www.bbc.co.uk")
+webDr$goBack()
+webDr$goForward()
+webDr$quit()
+
+## https://github.com/greenore/RSeleniumUtilities
+library(RSeleniumUtilities)
+RSeleniumUtilities::checkSelenium()
+webDr <- ieDriver()
+webDr <- firefoxDriver(use_profile=TRUE, profile_name="selenium")
+webDr <- chromeDriver(use_profile=TRUE, profile_name="selenium", internal_testing=TRUE)
+
+
+## Linear regression
+llply(split(mbase,mbase$Sess),function(x)lm(PL~Selection+HCap+Price,x))
+
+
+#'@ stopCluster(cl)
+
+x <- seq(as.Date('2011-01-01'), as.Date('2015-07-31'), by='months')
+y <- seq(min(mbase$PL),max(mbase$Stake), by=10000)
+labels <- date_format('%b')(x)
+breaks <- as.Date(sort(c(as.POSIXct(x), as.POSIXct(seq(min(mbase$Date), 
+          max(mbase$Date), by='months')), ymd('2015-08-01'))))
+labels <- c('', as.vector(rbind(labels, rep('', length(labels)))))
+
+ggplot(data=mbase, aes(x=x, y=y, shape=AHOU)) +
+  geom_line(aes(y = mbase$Stake, colour = 'Stake'), size=1.5) +
+  geom_line(aes(y = mbase$PL, colour = 'PL'), size=1.5) +
+  geom_point(size=2, fill='blue') + expand_limits(y=0) +         ## Set y range to include 0
+  scale_colour_hue(name='PL', l=30) +                          ## Set legend title use darker colors (lightness=30)
+  scale_shape_manual(name='PL', values=c(22,21)) +             ## Use points with a fill color
+  scale_shape_manual(values=c(22,21)) + xlab('Time of Day') + ylab('HK Dollars (HKD)') +
+  scale_x_date(labels = labels, breaks = breaks, limits=range(breaks)) + ## scale_x_date(labels = date_format("%b"),breaks = date_breaks("months")) +
+  ggtitle('Stakes and Profit & Lose') +                          ## Set title
+  theme_bw() + theme(legend.position=c(.7, .4))               ## Position legend inside this must go after theme_bw
+
+qplot(Stake, data=mbase, geom='density', fill=AHOU, alpha=I(.5), 
+      main='Turnover and P&L', xlab='Year in Month', 
+      ylab='HKD Amount') + scale_x_date(breaks=date_breaks('months'), labels = date_format("%b"))
+
+### http://statisticalrecipes.blogspot.com/2012/02/simulating-genetic-drift.html
+dtm <- factor(sapply(strsplit(as.character(mbase$Date),'-'),function(x) x[2]))
+dtm <- data.frame(month=mapvalues(dtm, sort(levels(dtm)),c('Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec')),
+                  mbase$Stake/10000, mbase$PL/10000); names(dtm) <- c('Month','Stake','PL')
+sdata <- data.frame(Date=factor(paste0(dtm$Month,'-',mbase$Sess)),dtm[-1]); rm(dtm)
+sdata <- ddply(sdata, .(Date), summarise, Stake=sum(Stake), PL=sum(PL))
+sdata[order(sdata$Date, decresing=FALSE),]
+
+## plot on same grid, each series colored differently -- 
+## good if the series have same scale
+ggplot(sim_data, aes(Month,'HKD 0000')) + geom_line(aes(colour = Series)) + 
+  scale_x_discrete(labels=c('Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec'))+
+  theme(axis.text.x=element_text(face="bold",colour="red",size=14))
+
+## ==================================================================================================================================
+## http://wenku.baidu.com/view/3574f639580216fc700afdfc.html
+## https://stat.ethz.ch/R-manual/R-devel/library/mgcv/html/gam.models.html
+## http://doc.qkzz.net/article/e6f33685-e220-4803-8c89-3228501b9412.htm
+
+
+
+
+
+

README.md

@@ -0,0 +1,2 @@
+# Betting-Strategy-and-Model-Validation
+Betting Strategy and Model Validation, I analyse the staking model of a sportsbook agency which follow bets from consultancy firm A

Testing efficiency of coding.Rmd

@@ -0,0 +1,70 @@
+---
+title: "Testing efficiency of coding"
+author: "Ryo®, Eng Lian Hu"
+date: "8/28/2015"
+output:
+  html_document:
+    fig_height: 3
+    fig_width: 5
+    highlight: haddock
+    theme: cerulean
+    toc: yes
+---
+
+This is an casting page of testing the efficiency of the coding for the research on `Betting Strategy and Model Validation`
+
+```{r load-packages}
+## Loading the packages
+if(!'devtools' %in% installed.packages()){
+  install.packages('devtools')}
+if(!'BBmisc' %in% installed.packages()){
+  install.packages('BBmisc')}
+
+suppressPackageStartupMessages(library('BBmisc'))
+pkgs <- c('devtools','RStudioAMI','zoo','chron','stringr','stringi','reshape','reshape2','data.table','sparkline','DT','plyr','dplyr','magrittr','parallel','foreach','memoise','manipulate','ggplot2','ggthemes','proto','extrafont','directlabels','PerformanceAnalytics','plotly','doMC','doParallel','BiocParallel','rvest','RSelenium','highlightHTML','knitr','rmarkdown','editR','scales','lubridate','tidyr','whisker','gtable','grid','gridExtra')
+suppressAll(lib(pkgs)); rm(pkgs)
+```
+
+```{r get-data-summary-table-2.1}
+nrow(do.call(rbind, llply(as.list(seq(2011,2015)), function(x) data.frame(Sess=x,read.csv(paste0(getwd(),'/datasets/',x,'.csv'))),.parallel=TRUE)))
+
+nrow(rbind_all(llply(as.list(seq(2011,2015)), function(x) data.frame(Sess=x,read.csv(paste0(getwd(),'/datasets/',x,'.csv'))),.parallel=TRUE)))
+
+system.time(do.call(rbind, llply(as.list(seq(2011,2015)), function(x) data.frame(Sess=x,read.csv(paste0(getwd(),'/datasets/',x,'.csv'))),.parallel=TRUE)))
+
+system.time(rbind_all(llply(as.list(seq(2011,2015)), function(x) data.frame(Sess=x,read.csv(paste0(getwd(),'/datasets/',x,'.csv'))),.parallel=TRUE)))
+```
+
+You can also embed plots, for example:
+
+```{r merge_all-dataframes-2.2}
+#'@ system.time(Reduce(function(x,y) {merge(x,y,all=TRUE)}, llply(list(df1,df1.sps,df1.pst),function(x) x[[1]])))
+
+#'@ system.time(merge_all(list(df1[[1]],df1.sps[[1]],df1.pst[[1]])))
+```
+
+Note that the `echo = FALSE` parameter was added to the code chunk to prevent printing of the R code that generated the plot.
+
+```{r}
+#'@ system.time(merge(socData, othData, all=TRUE))
+
+#'@ system.time(merge_all(list(socData, othData)))
+```
+
+
+
+  From the research, I learned from some articles which compare the efficiency of data measurement which will apply in future data analyse and data mining etc.
+
+- [Comparing performance of by, ddply and data.table](http://www.r-bloggers.com/transforming-subsets-of-data-in-r-with-by-ddply-and-data-table/)
+
+- [R高性能包介绍与并行运算](https://mp.weixin.qq.com/s?__biz=MzA3NDUxMjYzMA%3D%3D&mid=216065319&idx=1&sn=31af52816c7e8b937f15480c4d5f6e41&key=0acd51d81cb052bcbc420864d8003491eba2f4bbc722bf3a7bc7da0d59fefc64ea6fc32bdb33673eebd62f201cbc2190&ascene=7&uin=MjAwMTM4MjU0OA%3D%3D&devicetype=android-19&version=26020236&nettype=WIFI&pass_ticket=GdViEIR%2F5PLzVFnzLxc71K39ze4fb6VAwvFp1bhH3inbu5xBjyQ7BLEpDOrQhWZ1)
+
+- [A biased comparsion of JSON packages in R](https://rstudio-pubs-static.s3.amazonaws.com/31702_9c22e3d1a0c44968a4a1f9656f1800ab.html)
+
+- [Video how-to: Speed up R with C++ and Rcpp](http://www.computerworld.com/article/2961056/data-analytics/video-how-to-speed-up-r-with-c-plus-plus-and-rcpp-package.html)
+
+- [benchmarking logistic regression using glm.fit , bigglm, speedglm, glmnet, LiblineaR](http://stackoverflow.com/questions/19532651/benchmarking-logistic-regression-using-glm-fit-bigglm-speedglm-glmnet-libli)
+
+- [Dates and Times Made Easy with lubridate](http://www.jstatsoft.org/article/view/v040i03/v40i03.pdf)
+
+

custom.css

@@ -0,0 +1,13 @@
+/* http://www.w3schools.com/css/css_examples.asp */
+
+table {
+  max-width: 95%;
+  border: 1px solid #ccc;
+}
+th {
+  background-color: #0000FF;
+    color: #0000A0;
+}
+td {
+  background-color: #00FFFF;
+}