N-gram Corpus Research Articles

Diachronic predictor of affective ratings of English words

This work considers implementation of a diachronic predictor of valence, arousal and dominance ratings of English words. The estimation of affective ratings is based on data on word co-occurrence statistics in the large diachronic Google Books Ngram corpus. Affective ratings from the NRC VAD dictionary are used as target values for training. When tested on synchronic data, the obtained Pearson‘s correlation coefficients between human affective ratings and their machine ratings are 0.843, 0.779 and 0.792 for valence, aroused and dominance, respectively. We also provide a detailed analysis of the accuracy of the predictor on diachronic data. The main result of the work is creation of a diachronic affective dictionary of English words. Several examples are considered that illustrate jumps in the time series of affective ratings when a word gains a new meaning. This indicates that changes in affective ratings can serve as markers of lexical-semantic changes.

Have color representations in books changed over the past 200 years? An empirical analysis based on the Google Books Ngram corpus

AbstractColor significantly shapes our perceptions of the world, and extensive scientific research has explored various aspects of color, including color linguistics, aesthetics, color and marketing, color therapy, and color psychology. Notably, there has been an increasing interest in understanding the psychological domain of color in recent years. However, limited scholarly attention has been given to studying the changes in color representations and emotions over the decades. To address this gap, we utilized word embedding models of Google Books Ngrams corpus to examine the most strongly associated entities and traits with colors and their evolution in color emotions over 200 years. Analyzing three emotional dimensions—pleasure, arousal, and dominance (PAD), we found that red, green, white, black, orange, purple, brown, and pink were related to the exuberant type (+P + A + D) in PAD emotional space. In contrast, blue was linked to the relaxed type (+P−A + D), and yellow constituted the anxious type (−P + A−D). Among all basic English color terms, the emotion‐driven perceptions associated with white, brown, and pink have significantly changed over time.

P&T-Inf: A Result Inference Method for Context-Sensitive Tasks in Crowdsourcing

Context-Sensitive Task (CST) is a complex task type in crowdsourcing, such as handwriting recognition, route plan, and audio transcription. The current result inference algorithms can perform well in simple crowdsourcing tasks, but cannot obtain high-quality inference results for CSTs. The conventional method to solve CSTs is to divide a CST into multiple independent simple subtasks for crowdsourcing, but this method ignores the context correlation among subtasks and reduces the quality of result inference. To solve this problem, we propose a result inference algorithm based on the Partially ordered set and Tree augmented naive Bayes Infer (P&T-Inf) for CSTs. Firstly, we screen the candidate results of context-sensitive tasks based on the partially ordered set. If there are parallel candidate sets, the conditional mutual information among subtasks containing context information in external knowledge (such as Google n-gram corpus, American Contemporary English corpus, etc.) will be calculated. Combined with the tree augmented naive (TAN) Bayes model, the maximum weighted spanning tree is used to model the dependencies among subtasks in each CST. We collect two crowdsourcing datasets of handwriting recognition tasks and audio transcription tasks from the real crowdsourcing platform. The experimental results show that our approach improves the quality of result inference in CSTs and reduces the time cost compared with the latest methods.

Aspectual pairs: Prefix vs. suffix way of formation

In linguistic theory, there is no common point of view on the question of whether verbs in aspectual pairs are in inflectional or derivational relations. At the same time, the prefix and suffix methods of forming aspectual pairs are contrasted in this respect. The publications (e.g. Janda Lyashevskaya 2011) pointed out the need to develop new quantitative approaches to this aspect of the text corpus. We propose two new approaches that compare the quantitative characteristics of aspectual pairs of both types. One approach is based on the Google Books Ngram corpus and analyzes the dynamics of the frequency of the use of words in pairs. The aspectual pairs from the databases created by Janda and Lyashevskaya are considered. For a numerical assessment of the degree of proximity of the frequency graphs, the Pearson correlation coefficients were used. The second approach introduces a numerical characteristic of the semantic proximity of verbs in pairs using modern computer methods. Semantic proximity of verbs is calculated as a standard cosine measure between vectors representing the compatibility of the considered verbs in the corpus. Several computer models and text corpora are considered. Both proposed approaches did not reveal significant numerical differences in semantic proximity between verbs in aspectual pairs with prefix and suffix pairing. This is in good agreement with the results of an early study by Janda and Lyashevskaya (2011). Together with the results of this work, our research shows that the suffixal and affixal ways of forming aspectual pairs have an equal status in terms of their classification as inflectional or derivational.

Measurement and prediction of the international reputation of Chinese women’s apparel and accessories

Chinese women’s apparel and accessories have been often mentioned by other countries in history and nowadays, and the underlying culture along with these apparel and accessories contributes to this developing world to some extent. The assessment of the international reputation for these kinds of apparel and accessories is important for better understanding the worldwide cultural exchange, globalization, social development and so on; however, it remains unknown. This study presents a work with the overall objective to measure and predict the international reputation of Chinese women’s apparel and accessories. To reach this objective, the study first proposes a data mining procedure to use online query volume data from the Google Ngram corpus to construct a reputation index in order to measure the international reputation of Chinese women’s apparel and accessories. Then, by establishing a series of socio-economic and political determinants associated with the international reputation of Chinese women’s apparel and accessories, several forms of support vector machine based models are developed, aiming to predict their international reputation. There are two contributions out of this work: (1) presenting the first ever large-scale quantitative examination of the trend and measurement of the international reputation of Chinese women's apparel and accessories, which provides a new method to identify the dynamic transition of sociocultural development; (2) constructing a quantitative model to predict the international reputation of Chinese women's apparel and accessories, which generalizes an implication to understand the other sociocultural object and its impact factors.

A Novel N-Gram-Based Image Classification Model and Its Applications in Diagnosing Thyroid Nodule and Retinal OCT Images.

Imbalanced classes and dimensional disasters are critical challenges in medical image classification. As a classical machine learning model, the n-gram model has shown excellent performance in addressing this issue in text classification. In this study, we proposed an algorithm to classify medical images by extracting their n-gram semantic features. This algorithm first converts an image classification problem to a text classification problem by building an n-gram corpus for an image. After that, the algorithm was based on the n-gram model to classify images. The algorithm was evaluated by two independent public datasets. The first experiment is to diagnose benign and malignant thyroid nodules. The best area under the curve (AUC) is 0.989. The second experiment is to diagnose the type of fundus lesion. The best result is that it correctly identified 86.667% of patients with dry age-related macular degeneration (AMD), 93.333% of patients with diabetic macular edema (DME), and 93.333% of normal individuals.

Neural Network Algorithm for Detection of New Word Meanings Denoting Named Entities

Lexical semantic change detection has been a rapidly developing field of science in recent years. Existed algorithms of lexical semantic change detection face difficulties when they are used to work with words denoting named entities. This paper proposes a method that allows one to reveal a word in a large corpus that started being used as a named entity, as well as to date the first usage of this word as a proper name. To solve this problem, firstly, we offer an algorithm that allows for detecting words in a large corpus denoting named entities. The recognizer is based on an analysis of co-occurrences with the most frequent words and was trained on data from the English subcorpus of the Google Books Ngram corpus. The achieved recognition accuracy of named entities is 98.44% on the test sample. Secondly, we test the possibility of applying the trained recognizer to diachronic data. The analysed cases show that the recognizer initially trained using the total bigram frequencies for a long time interval, at least for any frequent word, provides stable results for the annual frequency values. This can make the recognizer a good tool for language evolution studies, especially for detecting new meanings of words. The analysed cases show that the proposed method allows revealing new word meanings associated with named entities, as well as detecting genericized meaning of words that were earlier used as proper names.

A visual approach to interpreting the career of the network metaphor

Metaphor has had a prolific presence in sociology: as a subject of empirical study, as a tool for theorizing, and as ideas that live public lives beyond the ivory tower. Despite all this work, tracing the core dynamics of metaphors over time remains a persistent challenge. To address this problem, I propose a systematic analysis of metaphor anchored around four core questions. What do metaphors describe? How do metaphors change over time? What do metaphors do? And how do conventional metaphors emerge? I address these questions in the context of one persistent and prolific metaphor: the network metaphor. To do so, I offer a formal model of the career of metaphor by drawing on cognitive linguistics and sociological theories of public ideas and professional careers. Specifically, I integrate the Google Books Ngram corpus with computational techniques that leverage co-occurrence to visualize and interpret the career of the network metaphor with respect to four areas: its jurisdiction, temporality, ecology, and conventionality. I leverage the network metaphor as both a revelatory case for evaluating my model and a critical case for deepening our historical understanding of the network. My analysis lends validation to the argument that the network metaphor has achieved the status of a broad category for thinking about contemporary life. However, it also demonstrates that the network metaphor had a lively history prior to the twentieth century. This includes significant changes in how it was used in the first half of the nineteenth century, a career standing in for anatomical structures of the body in the mid-nineteenth century, and a predominant expression in simile form prior to the start of the twentieth century.

Humor meets morality: Joke generation based on moral judgement

Although humor enriches human lives, some jokes fail to amuse people because of a lack of morality. In this paper, we propose a mechanism capable of selecting humor based on moral criteria. To this end, we first construct a model based on an N-gram corpus and generate joke candidates using various template patterns. We then employ a moral judgement classifier based on a recurrent neural network and utilize the trained model for humor selection. The experimental results obtained from best–worst scaling demonstrate that this scheme is able to generate jokes with moral category labels. We confirmed that jokes about the classifier categorized as Loyalty and Authority, which are regarded as good in our study, are funnier than jokes about Fairness, Purity, Harm, Cheating, and Degradation. Although we did not confirm that there was a difference in the funny level between good and bad moral jokes, the results demonstrate that moral categories of humor can affect the funny level.

Generation of a dictionary of abstract/concrete words by a multilayer neural network

Large dictionaries of abstract/concrete words were compiled for several languages by interviewing native speakers. The Russian dictionary contains only one thousand words. This article proposes a new method for automatic generation of a large (tens of thousands of words) Russian dictionary of abstract/concrete words by using neural networks trained on the Google Books Ngram corpus. Estimates of the quality of the dictionary compiled using this method are obtained. The correlation coefficient between the estimates obtained using a neural network for the level of word concreteness and the estimates obtained based on the native speakers‵ responds is 0.778.

Constructing Arabic Language Resources from Google N-gram Dataset

The Google N-gram dataset contains millions of books in 22 different categories. Therefore, it is used widely as a language resource in natural language processing. The Arabic language has several resources, but most of them are small in size. As with any research, if training resources are large, then the validity of any research is higher. To build a new language resource manually by humans take years to accomplish, but using automatic methods can be done in a few days. Google N-gram dataset does not support the Arabic language yet. This will cause a challenge for researchers who use this dataset for extracting Arabic text. In this paper, an extraction method that can produce a lexicon and N-gram corpus from the English Google N-gram dataset is presented. The steps to build these recourses and to overcome Google N-gram dataset challenges are explained. The experiment results show the success of the proposed method in the automatic construction of the resources in a short time. This method will be useful to researchers in several research fields.

Historical Change in the Moral Foundations of Political Persuasion.

How have attempts at political persuasion changed over time? Using nine corpora dating back through 1789, containing over 7 million words of speech (1,666 documents in total), covering three different countries, plus the entire Google nGram corpus, we find that language relating to togetherness permanently crowded out language relating to duties and obligations in the persuasive speeches of politicians during the early 20th century. This shift is temporally predicted by a rise in Western nationalism and the mass movement of people from more rural to more urban areas and is unexplained by changes in language, private political speech, or nonmoral persuasion. We theorize that the emergence of the modern state in the 1920s had psychopolitical consequences for the ways that people understood and communicated their relationships with their government, which was then reflected in the levers of persuasion chosen by political elites.

The werther effect revisited: Do suicides in books predict actual suicides?

While most extant studies have suggested that the Werther effect can be found via news and electronic media, whether books would have a similarly independent role on suicide suggestion remains unknown. Based on the Google Books N-gram corpus, we constructed the first macro-level index of suicides via books (ISB) and examined the potential association between ISB and actual completed suicides in the United States between 1950 and 2000. To rule out the possible influence of alternative mass media, we construct relevant measures by using the New York Times corpus and the Internet Movie Database (IMDb). The results from Granger causality test reveals that suicides in nonfiction rather than fiction books significantly predict actual suicide rates during the second half of the twentieth century, extensively controlling for suicides in other media as well as context factors. With such novel application of massive content analysis using data of unprecedented size, we contribute to the extant literature not only by adding to our understandings of the important role of media in shaping the process of suicide contagion but also the dynamics of suicidal behavior and the relevant cultural trends that may not be adequately measured with traditional social science methods.

Intellectual games concept review in THE XIX – XXI century (Google book Ngram Corpus scientific materials base)

In the context of the development of information technology in the world, modern scientists have noted in children, youth and adults the processes of creative and intellectual degradation, the growth of functional illiteracy. In the 20th century, the value of general intelligence for each new generation was one more statistical “norm” than the previous one. Scientists have discovered that the activity of a man of the 21st century in his free time becomes more significant for the development of his intellect and creativity (gaining a state of "flow") than activity during working hours. The intellectual load during leisure time allows a person to maintain a high level of mental activity until the end of his life, to prevent brain degradation. In the scientific community, from the mid-19th century, the verbal definition of the term “intellectual games” is enclosed in different types of games. Drawing analogies between the concepts of “Quiz” and “intellectual games”, the authors of the article determine that they are the product of two opposites: mass, entertainment culture and intellectual culture, which is characteristic of a rational person, capable of endless development. The authors described a modern tool for working in the database of scientific materials Google Book Ngram Corpus. This database contains also materials concerning architecture, construction, machine-buliding fields of research. With its help, connotations of intellectual games in the history of their development were considered.

ETYMOLOGY IN THE EARTH SCIENCES: FROM ‘GEOLOGIA' TO ‘GEOSCIENCE’

The origin and usage through time of geologia, geognosy, geogony, oryctognosy, geology and geophysics, as characterised by their frequency of occurrence in the Google Books Ngram Corpus, is discussed. The English, French, German, Italian and Spanish corpuses used in this study have been normalised over the same timespan using the average frequencies of occurrence of the same set of ‘neutral’ words in each language (as advocated by Younes and Reips 2019). Use of the term geology is found to predate publication of James Hutton's Theory of the Earth in 1795 by about 100 years; geognosy, oryctognosy and geogony, much less commonly used, became established in the 1780s and began to fall out of use around 1820. The terms geologist, and geognost follow a similar pattern. The emergence of geophysics is a less familiar field: While the phrases physics of the Earth and physical geography can both be traced back to the early 1700s, geophysics only began to be used in the early 1800s and did not really become common until about 1860; geophysicist becomes common in German after 1860, but more generally after 1880. The first geophysics-related publications were bulletins from magnetic and seismic observatories and its first dedicated journal, Beiträge zur Geophysik, began publication in 1887, eighty years after the formation of The Geological Society of London. The tems earth science and geoscience, popular today, have steadily increased in their usage since being introduced in the 1880s and 1930s respectively.

Analysis of the probability distribution of time series of word frequencies

Knowledge of probability distribution of frequency of words and phrases is important in solving many practical problems such as estimation of semantic similarity between words, detection of semantic changes and others. It is assumed by many researchers that frequencies of words obey the Poisson law. However, there is much evidence that the Poisson distribution describes empirical data unsatisfactory. The analysis of the probability law in this case is greatly complicated by the fact that series of frequencies in most cases are non-stationary. This paper discusses the distribution law of time series of word frequencies based on the Google Books Ngram corpus data. It is shown that the correlation between the first moments of the frequencies differs from that expected in the assumption of the Poisson distribution. In particular, anomalously high values of frequency dispersion are observed for words with high frequency. To check the significance of deviations from the Poisson law, statistical modeling of frequency series was performed.

Dictionary of Abstract and Concrete Words of the Russian Language: A Methodology for Creation and Application

The paper describes the first stage of a project on creating an electronic dictionary with numerical estimates of the degree of abstractness and concreteness of Russian words. Our approach is to integrate data obtained from several different sources: text corpora, psycholinguistic experiments, published dictionaries, markers of abstractness (certain suffixes) and a translation of a similar dictionary for English. The article describes in detail the method for obtaining data. We provide correlation coefficients calculated using different methods. We pay special attention to cases of inconsistent results obtained by different methods. The statistical model behind the experimental data is discussed. The results of experiments with the Google Books Ngram corpus on the coexistence of concrete words are given. Possible applications of the dictionary are demonstrated on the example of the frequency of using the dictionary in Russian textbooks for high schools.

Gender associations in the twentieth-century English-language literature

Gender associations may be partly learned from print media, including literature. Using Google Books Ngram corpus, we explored the depiction of male and female characters in the twentieth-century English-language fiction. By analyzing adjective-noun bigrams, we examined adjectives used in association with “man”, “woman”, “boy”, and “girl”. Men were described in more positive terms than women. Girls were depicted more positively than boys at the beginning of the twentieth century, but the tendency reversed in the middle of the century. Boys were described in more masculine terms than girls; however, men were described in similarly masculine adjectives as women. Despite limitations of interpretability of the results, the study presents a possible approach of exploring past characterization of the two genders.

The natural selection of words: Finding the features of fitness.

We introduce a dataset for studying the evolution of words, constructed from WordNet and the Google Books Ngram Corpus. The dataset tracks the evolution of 4,000 synonym sets (synsets), containing 9,000 English words, from 1800 AD to 2000 AD. We present a supervised learning algorithm that is able to predict the future leader of a synset: the word in the synset that will have the highest frequency. The algorithm uses features based on a word’s length, the characters in the word, and the historical frequencies of the word. It can predict change of leadership (including the identity of the new leader) fifty years in the future, with an F-score considerably above random guessing. Analysis of the learned models provides insight into the causes of change in the leader of a synset. The algorithm confirms observations linguists have made, such as the trend to replace the -ise suffix with -ize, the rivalry between the -ity and -ness suffixes, and the struggle between economy (shorter words are easier to remember and to write) and clarity (longer words are more distinctive and less likely to be confused with one another). The results indicate that integration of the Google Books Ngram Corpus with WordNet has significant potential for improving our understanding of how language evolves.

Combining semantic and term frequency similarities for text clustering

A key challenge for document clustering consists in finding a proper similarity measure for text documents that enables the generation of cohesive groups. Measures based on the classic bag-of-words model take into account solely the presence (and frequency) of words in documents. In doing so, semantically similar documents which use different vocabularies may end up in different clusters. For this reason, semantic similarity measures that use external knowledge, such as word n-gram corpora or thesauri, have been proposed in the literature. In this paper, the Frequency Google Tri-gram Measure is proposed to assess similarity between documents based on the frequencies of terms in the compared documents as well as the Google n-gram corpus as an additional semantic similarity source. Clustering algorithms are applied to several real datasets in order to experimentally evaluate the quality of the clusters obtained with the proposed measure and compare it with a number of state-of-the-art measures from the literature. The experimental results demonstrate that the proposed measure improves significantly the quality of document clustering, based on statistical tests. We further demonstrate that clustering results combining bag-of-words and semantic similarity are superior to those obtained with either approach independently.