1,720,998 research outputs found
Exploring the Semantic Meaning of Constructs that Lead to Human Decisions
Get PDFThis study examines automated approaches to discovering behavioral knowledge that are encoded as constructs in social and behavioral science disciplines. To date, constructs relationships are ordinarily revealed through laborious psychometric methods, but this study has shown that it is possible to extract these relationships through automated computational approaches. By building on text similarity measures from prior literature, we are able to predict construct relationships through construct name, definition and items. The predicted relationships were woven into an interlock system to demonstrate construct interplays, even though they have not been studied. The construct interlock could be seen as a theory map to understand human decision-making. Two use cases were presented to demonstrate the efficacy of the proposed measures: measuring the root constructs in UTAUT and visualizing network of construct perceived usefulness. The encouraging results showed that the proposed measures could dramatically expedite theory development, at the same time also expedite progression of human science
A machine learning framework for automated text categorization
Get PDFThis dissertation describes a machine learning framework for the development of an automated text categorization system for real-life problems. Conference paper classification will be used as a case study of a life text categorization problem. Unlike documents in benchmark collections, text documents such as conference papers tend to be rather heterogeneous having a rich structure with variable length documents where each category consists of a variable number of documents
Exploring the Semantic Meaning of Constructs that Lead to Human Decisions
Get PDFThis study examines automated approaches to discovering behavioral knowledge that are encoded as constructs in social and behavioral science disciplines. To date, constructs relationships are ordinarily revealed through laborious psychometric methods, but this study has shown that it is possible to extract these relationships through automated computational approaches. By building on text similarity measures from prior literature, we are able to predict construct relationships through construct name, definition and items. The predicted relationships were woven into an interlock system to demonstrate construct interplays, even though they have not been studied. The construct interlock could be seen as a theory map to understand human decision-making. Two use cases were presented to demonstrate the efficacy of the proposed measures: measuring the root constructs in UTAUT and visualizing network of construct perceived usefulness. The encouraging results showed that the proposed measures could dramatically expedite theory development, at the same time also expedite progression of human science
A Framework to Predict Software “Quality in Use” from Software Reviews
Get PDFSoftware reviews are verified to be a good source of users’ experience.
The software “quality in use” concerns meeting users’ needs. Current
software quality models such as McCall and Boehm, are built to support software
development process, rather than users perspectives. In this paper, opinion
mining is used to extract and summarize software “quality in use” from software
reviews. A framework to detect software “quality in use” as defined by the
ISO/IEC 25010 standard is presented here. The framework employs opinionfeature
double propagation to expand predefined lists of software “quality in
use” features to domain specific features. Clustering is used to learn software
feature “quality in use” characteristics group. A preliminary result of extracted
software features shows promising results in this direction
An empirical study of feature selection for text categorization based on term weightage
Get PDFThis paper proposes a local feature selection (FS) measure namely, Categorical Descriptor Term (CTD) for text categorization. It is derived based on classic term weighting scheme, TFIDF. The method explicitly chooses feature set for each category by only selecting set of terms from relevant category. Although past literatures have suggested that the use of features from irrelevant categories can improve the measure of text categorization, we believe that by incorporating only relevant feature can be highly effective. The experimental comparison is carried out between CTD and five wellknown feature selection measures: Information Gain, Chi-Square, Correlation Coefficient, Odd Ratio and GSS Coefficient. The results also show that our proposed method can perform comparatively well with other FS measures, especially on collection with highly overlapped topics
Joint Distance and Information Content Word Similarity Measure
Get PDFMeasuring semantic similarity between words is very important to many applications related to information retrieval and natural language processing. In the paper, we have discovered that word similarity metrics suffer from the drawback of obtaining equal similarities of two words, if they have the same path and depth values in WordNet. Likewise information content methods which depend on word probability of a corpus tend to posture the same drawback. This paper proposes a new hybrid semantic similarity to overcome the drawbacks by exploiting advantages of Li and Lin methods. On a benchmark set of human judgments on Miller Charles and Rubenstein Goodenough data sets, the proposed approach outperforms existing methods in distance and information content based methods
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Measuring Software Quality in Use: State-ofthe-Art and Research Challenges
Get PDFSoftware quality in use comprises quality
from the user’s perspective. It has gained
its importance in e-government applications,
mobile-based applications, embedded
systems, and even business process development.
Users’ decisions on software
acquisitions are often ad hoc or based on
preference due to difficulty in quantitatively
measuring software quality in use. But,
why is quality-in-use measurement difficult?
Although there are many software quality
models, to the authors’ knowledge no works
survey the challenges related to software
quality-in-use measurement. This article
has two main contributions: 1) it identifies
and explains major issues and challenges
in measuring software quality in use in the
context of the ISO SQuaRE series and related
software quality models and highlights open
research areas; and 2) it sheds light on a
research direction that can be used to predict
software quality in use. In short, the qualityin-
use measurement issues are related to the
complexity of the current standard models
and the limitations and incompleteness of
the customized software quality models. A
sentiment analysis of software reviews is
proposed to deal with these issues
An empirical study on CO2 emissions in ASEAN countries
Get PDFThis paper proposes a local feature selection (FS) measure namely, Categorical Descriptor Term (CTD) for text categorization. It is derived based on classic term weighting scheme, TFIDF. The method explicitly chooses feature set for each category by only selecting set of terms from relevant category. Although past literatures have suggested that the use of features from irrelevant categories can improve the measure of text categorization, we believe that by incorporating only
relevant feature can be highly effective. The experimental
comparison is carried out between CTD and five wellknown
feature selection measures: Information Gain, Chi-Square, Correlation Coefficient, Odd Ratio and GSS Coefficient. The results also show that our proposed method can perform comparatively well with other FS measures, especially on collection with highly overlapped topics
- …
