我尝试使用ngram（bigram，trigram等）提取频率。但是在最终结果中我采用单项的频率。为什么会这样？我有什么必须在代码中修复。
以下是我使用的数据：
df <- structure(list(text = c("the discipline of phenomenology is defined by its domain of study its methods and its main results ",
"phenomenology studies structures of conscious experience as experienced from the first person point of view along with relevant conditions of experience the central structure of an experience is its intentionality the way it is directed through its content or meaning toward a certain object in the world ",
"we all experience various types of experience including perception imagination thought emotion desire volition and action thus the domain of phenomenology is the range of experiences including these types among others experience includes not only relatively passive experience as in vision or hearing but also active experience as in walking or hammering a nail or kicking a ball the range will be specific to each species of being that enjoys consciousness our focus is on our own human experience not all conscious beings will or will be able to practice phenomenology as we do ",
"conscious experiences have a unique feature we experience them we live through them or perform them other things in the world we may observe and engage but we do not experience them in the sense of living through or performing them this experiential or first person feature — that of being experienced — is an essential part of the nature or structure of conscious experience as we say “i see think desire do …” this feature is both a phenomenological and an ontological feature of each experience it is part of what it is for the experience to be experienced phenomenological and part of what it is for the experience to be ontological ",
"how shall we study conscious experience we reflect on various types of experiences just as we experience them that is to say we proceed from the first person point of view however we do not normally characterize an experience at the time we are performing it in many cases we do not have that capability a state of intense anger or fear for example consumes all of one s psychic focus at the time rather we acquire a background of having lived through a given type of experience and we look to our familiarity with that type of experience hearing a song seeing a sunset thinking about love intending to jump a hurdle the practice of phenomenology assumes such familiarity with the type of experiences to be characterized importantly also it is types of experience that phenomenology pursues rather than a particular fleeting experience — unless its type is what interests us ",
"classical phenomenologists practiced some three distinguishable methods we describe a type of experience just as we find it in our own past experience thus husserl and merleau ponty spoke of pure description of lived experience we interpret a type of experience by relating it to relevant features of context in this vein heidegger and his followers spoke of hermeneutics the art of interpretation in context especially social and linguistic context we analyze the form of a type of experience in the end all the classical phenomenologists practiced analysis of experience factoring out notable features for further elaboration ",
"these traditional methods have been ramified in recent decades expanding the methods available to phenomenology thus in a logico semantic model of phenomenology we specify the truth conditions for a type of thinking say where i think that dogs chase cats or the satisfaction conditions for a type of intention say where i intend or will to jump that hurdle in the experimental paradigm of cognitive neuroscience we design empirical experiments that tend to confirm or refute aspects of experience say where a brain scan shows electrochemical activity in a specific region of the brain thought to subserve a type of vision or emotion or motor control this style of “neurophenomenology” assumes that conscious experience is grounded in neural activity in embodied action in appropriate surroundings — mixing pure phenomenology with biological and physical science in a way that was not wholly congenial to traditional phenomenologists ",
"what makes an experience conscious is a certain awareness one has of the experience while living through or performing it this form of inner awareness has been a topic of considerable debate centuries after the issue arose with locke s notion of self consciousness on the heels of descartes sense of consciousness conscience co knowledge does this awareness of experience consist in a kind of inner observation of the experience as if one were doing two things at once brentano argued no is it a higher order perception of one s mind s operation or is it a higher order thought about one s mental activity recent theorists have proposed both or is it a different form of inherent structure sartre took this line drawing on brentano and husserl these issues are beyond the scope of this article but notice that these results of phenomenological analysis shape the characterization of the domain of study and the methodology appropriate to the domain for awareness of experienc... <truncated>
"conscious experience is the starting point of phenomenology but experience shades off into less overtly conscious phenomena as husserl and others stressed we are only vaguely aware of things in the margin or periphery of attention and we are only implicitly aware of the wider horizon of things in the world around us moreover as heidegger stressed in practical activities like walking along or hammering a nail or speaking our native tongue we are not explicitly conscious of our habitual patterns of action furthermore as psychoanalysts have stressed much of our intentional mental activity is not conscious at all but may become conscious in the process of therapy or interrogation as we come to realize how we feel or think about something we should allow then that the domain of phenomenology — our own experience — spreads out from conscious experience into semi conscious and even unconscious mental activity along with relevant background conditions implicitly invoked in ... <truncated>
)), .Names = "text", row.names = c(NA, -9L), class = "data.frame")
这是我用来获取bigram和trigram的代码：
library("tm")
library(slam)
df2 <- df$text
#use tm package
review_source <- VectorSource(df2)
corpus <- Corpus(review_source)
corpus <- tm_map(corpus, stripWhitespace)
corpus <- tm_map(corpus, removeWords, stopwords("english"))
# Functions
BigramTokenizer <- function(x) {RWeka::NGramTokenizer(x, RWeka::Weka_control(min=2, max=2))}
# Bigrams
options(mc.cores=1)
dtm.docs.2g <- DocumentTermMatrix(docs.s, control=list(tokenize=BigramTokenizer))
# To get the bigram dist, we use the slam package for ops with simple triplet mat
sums.2g <- colapply_simple_triplet_matrix(dtm.docs.2g,FUN=sum)
sums.2g <- sort(sums.2g, decreasing=T)
当我输入控制台时，我预期的结果sums.2g是2个单词短语的结果，我只使用一个频率。这里是我提到的结果的示例输出。
sums.2g 体验有意识的现象学类型 44 15 13 10 域名第一人称 6 5 5 5
答案 0 :(得分：2)
你有
class(corpus)
# [1] "SimpleCorpus" "Corpus"
?DocumentTermMatrix解释了它如何处理<{p}的control参数
这对于SimpleCorpus来说是不同的。在这种情况下，所有选项都是 一次性以固定顺序处理以提高性能。它 总是使用Boost Tokenizer（通过Rcpp）并且不需要自定义 作为选项参数。
使用corpus <- VCorpus(review_source)。