In f0nzie/petroleum.ml: Machine Learning Applications in Machine Learning

# read the user modified Excel file with PE discipline assignment
devres_loc <- "./inst/results"
xls.nodups.papers.4 <- readxl::read_xlsx(paste(devres_loc, "ml_nodups_papers_4.xlsx", sep = "/"), sheet = 1)
xls.nodups.papers.4

library(dplyr)

nodups.papers.ren <- xls.nodups.papers.4 %>% 
    select(-c(X__1, rec)) %>% 
    rename(discipline = V2, algorithm = V3)

nodups.papers.ren

Papers manually selected

# find one paper at a time
grep("172610", nodups.papers.ren$id, value = TRUE)

papers.1 <- c("172610",
                "69624",
                 "67260",
                 "84441",
                 "181683",
                 "150314",
                 "143842",
                 "68163",
                 "90259")

find_paper <- function(x) {
    grep(x, nodups.papers.ren$id, value = TRUE)
}

# res.1 <- matrix(unlist(sapply(papers.1, find_paper)),  nrow=10, byrow=T)
res.1 <- sapply(papers.1, find_paper)

# https://stackoverflow.com/a/4227504/5270873
data.frame(sapply(res.1,c))

# https://stackoverflow.com/a/4227483/5270873
# the problem is that values repeat in the 2nd column.
do.call(rbind.data.frame, res.1)

data.frame(Reduce(rbind, res.1))
# number of columns of result is not a multiple of vector length (arg 2)
# some row.names duplicated: 3,4,5,6,7,8,9 --> row.names NOT used

# this one works
# it adds a counter to the row.names with the paper that has two different types
# https://stackoverflow.com/a/29916958/5270873
do.call(rbind, lapply(res.1, data.frame, stringsAsFactors=FALSE))

f = function(x) function(i) unlist(lapply(x, `[[`, i), use.names=FALSE)
as.data.frame(Map(f(res.1), names(res.1[[1]])))

find_paper("150314")

find_paper <- function(x) {
    ret <- grep(x, nodups.papers.ren$id, value = TRUE)
    if (identical(ret, character(0))) ret <- NA
    return(ret)

}

find_paper("27561")

paper_list <- "
            172610
            69624
             67260
             84441
             181683
             150314
             143842
             68163
             90259
            27561
            18158
            52959
            78334
            170683
            94357
            85650
            17413
            69405
            87008
            27905
            30556
            130095
            170683
            99667
            170660
            167897
            101474
            102492
            86467
            100131
            122186
            83974
            133831
            99882
            90720
            94357
            90404
            136373
            90372
            128636
            170866
            136373
            112223
            56433
            19619
            29220
            46206
            69405
"

papers_list.df <- read.table(text = paper_list, stringsAsFactors = FALSE, 
                       colClasses = "character", col.names = "paper")

find_paper <- function(x) {
    ret <- grep(x, nodups.papers.ren$id, value = TRUE)
    if (identical(ret, character(0))) ret <- NA
    return(ret)

}


papers.2 <- read.table(text = paper_list, stringsAsFactors = FALSE, 
                       colClasses = "character", col.names = "paper")
# papers.2
# grep("90259", nodups.papers.ren$id, value = TRUE)

res.2 <- sapply(papers.2$paper, find_paper)


# papers not found are not shown
# https://stackoverflow.com/a/29916958/5270873
df <- do.call(rbind, lapply(res.2, data.frame, stringsAsFactors=FALSE))
names(df)[1] <- "paper"
df$search <- row.names(df)
df <- df[, c(2,1)]
row.names(df) <- NULL
df

10 out 34 papers were found in the paper search (746). What about trying different kind of search?

# build the new find_papers() function

#' Find papers in a paper_id column vector
#' 
#' The function looks for the papers in the input vector x in the given dataframe 
#' column corresponding to the paper_id
#' @param x input vector of paper to be searched
#' @param df_papers_id paper_id in the dataframe to be searched
#' @export
find_papers <- function(x, df_papers_id) {
    .find_paper <- function(x, df_papers_id) {
    ret <- grep(x, df_papers_id, value = TRUE)
    if (identical(ret, character(0))) ret <- NA
    return(ret)
    }

    df <- do.call(rbind, lapply(sapply(x, .find_paper, df_papers_id), 
                            data.frame, stringsAsFactors=FALSE))
    names(df)[1] <- "paper"
    df$search <- row.names(df)
    df <- df[, c(2, 1)]
    row.names(df) <- NULL
    df
}

pap <- c("90259", "27561", "1706831", "181683", "150314")

find_papers(pap, nodups.papers.ren$id)

find_papers(papers.2$paper, nodups.papers.ren$id)

What keyword did we miss when we were searching the papers? Why papers that have been downloaded were not selected in the search?

Looking for papers that have neural network in the title or body

library(petro.One)

major <- c("neural network")
minor <- c("reservoir", "production", "logging", "completion", "intervention",
           "drilling", "geology", "seismic", "petrophysics", "geophysics", 
           "economics")

# the returning data structure is a a list
# the list contains two dataframes: one for the keywords and a second for the papers
prod.li <- join_keywords(major, minor, get_papers = TRUE, sleep = 3)
prod.li

found.df <- find_papers(papers_list.df$paper, prod.li$papers$paper_id)

summary(found.df)

str(found.df)

found.df[which(is.na(found.df$paper )),]

SPE 29220 Artificial Neural Network As A Valuable Tool For Petroleum Engineers

library(petro.One)

major <- c("Neural Network")

# the returning data structure is a a list
# the list contains two dataframes: one for the keywords and a second for the papers
prod.li2 <- join_keywords(major, get_papers = TRUE, sleep = 3)
prod.li2

found.df <- find_papers(papers_list.df$paper, prod.li2$papers$paper_id)
found.df[which(is.na(found.df$paper )),]

write.csv( prod.li2$papers, file = "neural_network.csv")

library(petro.One)

major <- c("neural network", "neural-network", "NEURAL NETWORK", "Neural Networks",
           "neuro", "neural", 
           "machine learning", "MACHINE LEARNING",
           "pattern recognition",
           "artificial neural network", 
           "data driven", "data-driven", "data mining", "analytics",
           "artificial intelligence", "artificial intelligent", 
           "intelligent system", "cognitive",
           "dataset", "datasets", "data set", "data sets")

# the returning data structure is a a list
# the list contains two dataframes: one for the keywords and a second for the papers
prod.li3 <- join_keywords(major, get_papers = TRUE, sleep = 3)
prod.li3

found.df <- find_papers(papers_list.df$paper, prod.li3$papers$paper_id)
found.df[which(is.na(found.df$paper )),]

f0nzie/petroleum.ml documentation built on May 12, 2019, 6:22 p.m.