2-3. Previous Research

Those studies above clearly suggest that “Sentimental Analysis” contains a lot of aspects which have been developed in various fields. Additionally, it would be beneficial for us to systematize the whole elements which are introduced the previous part.

It must be safe to say that the direct ancestors of “Sentimental Analysis” are “Content Analysis”, “Statistics” and “Computer Science” because the new style uses every component those preceding research have established for a long time. If we show the relationship in a figure, that would be like Figure.5 below. The illustration explains how different territories of linguistic synergize with each other as well as how Sentimental Analysis is constructed academically.

 In “Foundations of behavioural research (1986)”, Kerlinger stated that “Content Analysis” is characterized as “a method of studying and analyzing communication in a systematic, objective, and quantitative manner for the purpose of measuring variables (Content Analysis A method, p.2).” On the other hand, “Quantitative Content Analysis” is a part of “Content Analysis” and it especially features the aspect of “Statistics” more to attain the objective. Zhang and Wildemuth introduced the three definitions for “Quantitative Content Analysis”;

• “a research method for the subjective interpretation of the content of text data through the systematic classification process of coding and identifying themes or patterns”

• “an approach of empirical, methodological controlled analysis of texts within their context of communication, following content analytic rules and step by step models, without rash quantification”

• “any qualitative data reduction and sense-making effort that takes a volume of qualitative material and attempts to identify core consistencies and meanings” (Zhang & Barbara, p. 1)

Moreover, they also indicate the strong tie between the two approaches saying; “The quantitative approach produces numbers that can be manipulated with various statistical methods (Zhang & Barbara, p. 2).” Therefore, it would be rational inference to formulate that the combination of “Content Analysis” and “Statistics” creates “Quantitative Content Analysis”. The same thing can be said about the rapport between “Content Analysis” and “Computer Science” which creates “Computer Content Analysis”. “Computer Content Analysis” is a computer-assisted evaluation method that treats text or its contents (Alexa, 1997, p. 5). This modern technics has been evolved largely due to the abundant electronic text and databases from large variety of sources which are available for researchers who dedicate their time and effort to mine the information for text structures. One notable trait that machine-readable text has is that it is recognizable by application so that many data the target materials have can easily be collected and sorted in order to analyze them. As stated in Alexa’s report in 1997, computer systems such as "General Inquirer" which is an IBM 7090 program system that was developed at Harvard in the spring of 1961 by Philip J. Stone and his colleague and "WORDS" proposed in 1969 by H.P. Iker and N.I. Harway contributed to the establishment of “Computer Content Analysis” (Alexa, 1997, p. 5). In addition to those software for massive “Content Analysis”, Alexa pointed out that DeWeese’s proposals and technics shown in his two theses during late 70’s (DeWeese, 1976 & 1977) may be regarded as path-breaking works of the inception of “Computer Content Analysis”. The progress above clearly suggests that “Computer Content Analysis” consists of “Content Analysis” and “Computer Content Analysis”. Though there might be some possibilities about the involvement of “Statistics” in the aseessment, it is a logical determination to omit “Statistics” from the components of “Computer Content Analysis” because the main purpose of the computerized approach focuses on the exploration of the text structures, rather than numerical facts or data.

The third academic territory given birth of as a result of theoretical fusion is “Statistical Machine Learning (a.k.a: Statistical learning theory)” which is defined as a unification of statistics and the computational sciences.  “Statistical Machine Learning” generally means the self-educated computer system for “the automated detection of meaningful patterns in data (Shai, 2014, 9. 7).” Recently, this kind of technology has widely used for creating various applications we use every day. For example, search engines represented by Google, Yahoo, MSN, and portal sites like that are made up based on this methodical approach. Furthermore, “Statistical Machine Learning” has a unique feature that can execute any complicated tasks by its ability to “learn” and “adapt”. Those capacity is able to be achieved only by “Statistical Machine Learning”. This is because that it is impossible for human engineers to program a system to deal with every assignment which needs an explicit and detailed specification.  Thus, the mechanics also plays a crucial role in Artificial Intelligence which performs operations analogous to learning and decision making in humans.  As the name suggests, “Statistical Machine Learning” has deep relationship with “Statistics”. In a book “An Introduction to Statistical Learning”, the authors say that “statistical learning (“Statistical Machine Learning”) has emerged as a new subfield in statistics, focused on supervised and unsupervised modeling and prediction (James, 2013).”As shown above, “Sentimental Analysis” is a trinity complex including the three basic elements; “Content Analysis”, “Statistics” and “Computer Science”. Moreover, it is hugely affected by the derivatives like “Quantitative Content Analysis”, “Computer Content Analysis” and “Statistical Machine Learning”. In other word, “Sentimental Analysis” represents a fruits of long-term research from various fields.

In summary, “Content Analysis”, “Statistics” and “Computer Science” are distinct stand-alone area respectively. Nevertheless, by cooperating with each other, they successfully invented different genres. As a result, the movement make it possible to emerge the three subsidiaries and one next-generation technology; “Quantitative Content Analysis”, “Computer Content Analysis”, “Statistical Machine Learning” and “Sentimental analysis”. As for the three subsidiaries, it might be difficult to set a boundary between them strictly because all of them are influenced by the three basic parent elements to some extent. However, in this thesis, it would be convenient to treat them as independent existences for the purpose to grasp the latent functions of “Sentimental Analysis” systematically.