The existence of a latent temporal pattern in keywords’ occurrences is explored by means of (lexical) correspondence analysis and clusters of keywords portraying similar temporal patterns are identified by functional (textual) data analysis and model-based curve clustering.
The analyses reveal a definite time dimension in topics and show that much of the History of Statistics may be gleaned by simply reading the titles of papers through an explorative correspondence analysis.
Increasingly language is being exploited to gain insights into human behavior and social interactions.
Speech and natural language processing has been applied, for example, to infer opinions from spoken utterances and text [4, 12], social dominance and stance from dialogues [23, 22], speaker’s emotion from speech , links between evolution of language and sexual selection , summary from dialogues , social interaction style from conversations , and sexual interaction styles – friendly, flirtatious, awkward, and assertive – in speed-dating conversations .
In this study a set of keywords occurred in the titles of papers published in the period 1888–2012 by the Journal of the American Statistical Association and its predecessors are examined over time in order to retrieve those which appeared in the past and which are today the research fields covered by Statistics, from the viewpoints of both methods and application domains.
In the same vein, this paper examines the task of inferring social relationships between conversants engaged in everyday conversations, a task that has not been examined in the literature before.
This task is motivated by the need to quantify social interactions with one’s social relationships – broadly referred to as social engagement in epidemiological literature – in older adults.
In addition to studying the task, we demonstrate the feasibility of characterizing social engagement from everyday telephone conversations using spoken language processing techniques.
The task of classifying social relationships has not been performed before mostly due to the lack of a naturalistic collection of everyday conversations.