Development and Validation of a Self-Concept Scale for College Students Using Comics Superhero Characters

Self-concept refers to the domain of self-descriptions that have self-evaluative connotation. Though many researchers embarked in the study of self-concept, and some even developed tests that measured self-concept, majority of these instruments had methodological and theoretical problems due to lack of systematic instrument development and presentation. The objective of the study is to develop a reliable and valid alternative approach to measuring the self in a semi-structured undisguised comics-type test that directly accounts for the way college students consider their choices of superheroes’ traits that characterizes their own. A preliminary survey on self-concept, in a form of open-ended statements was conducted to five hundred ninety-eight (598) college students of selected schools in Manila and Bulacan to know how college students see themselves indicatory of their self-concepts. Results of which, were collated to form the preliminary form. The preliminary form of the SCSS was administered to five hundred ninety-five (595) college students of different universities and colleges. Eighty-eight (88) items under eight (8) components were subjected to item analysis by identifying factors through a series of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Descriptive results were also calculated, as well as the exact reliability coefficient through split-half and Cronbach’s alpha. For the validity, content analysis was applied using two groups of experts who ascertain the suitability of each item in terms of content, relevance, clarity, appropriateness and their representations. They include three (3) experts who have a long experienced in comics industry and another three (3) experts in the field of college students’ self-concept formation. From the total of 88 items, 30 items were eliminated. However, the items that constitute the final form of the SCSS was concentrated into 55 items under six (6) factors upon post-analysis consideration. Statistical analysis revealed that the experts’ ratings were consistent and has high reliability with a generated r value of.894. The SCSS final form was administered to 809 respondents following the same procedures that were used for the preliminary form. The test scores were subjected to reliability facility, such as Alpha Coefficient and Split-Half, computing the reliability coefficients of the final form. Validity was established through convergent analysis, tested in a sample of 419 respondents who took the Tennessee Self-Concept Scale (TSCS: 2) Adult Form. The test was found to have high reliability with r =.792.

Nature Inspired Computation Techniques in Economic and Financial Modeling

Economic and financial modeling is the task of making mathematical descriptions for complex phenomena occurring in financial markets. It is proven fact that traders and decision makers in financial markets make better decisions if they have sound indications about future events that can have potential impact on the financial instruments. This prior knowledge is very important for business risk analysis, credit scoring, competitive market analysis, portfolio management and financial forecasting etc. Due to incomplete and imprecise knowledge of influencing factors and the implicit nonlinearities, financial decision making remains complex and challenging. Most of the natural phenomena are dynamic systems with inherent complexities. Algorithms based on these natural phenomena are best suited for solving intricate and multidimensional problems. Nature inspired algorithms are stochastic search algorithms which get their inspiration from nature. These natural phenomena include fish schooling, bird flocking, animal herding, biological evolution, natural selection, etc. These algorithms can be applied to financial and economic modeling tasks due to their ability to solve and perform better in complex situations. In this dissertation nature inspired algorithms are applied on economic and financial modeling problems for improving the performance measures of traditional econometric and financial models. The contributions in this thesis are summarized as follows: .Contribution 1: A two phase method using Genetic Algorithm and Particle Swarm optimization is formulated for fuzzy time series forecasting. This method uses twofactor, kth order fuzzy logical relationship groups and a weighted forecasting formula for prediction of stock market index of Taiwan Futures Exchange (TAIFEX). With this new approach, it is shown that proposed method has better convergence rate, better optimization, and lower predictive modeling. Contribution 2: A hybrid Genetic Algorithm and Particle Swarm optimization based fuzzy time series forecasting algorithm is proposed. This method uses Genetic Algorithm and Particle Swarm Optimization in parallel, working on the principle of elitism on individuals. This method is employed on stock market index forecasting of TAIFEX index and Karachi Stock Exchange (KSE-100) index. Contribution 3: Quantum computing is an emerging paradigm having potential applications in all domains of computing. A novel approach using Quantum Evolutionary algorithm for fuzzy time series forecasting is proposed. Quantum Evolutionary algorithm is used along with fuzzy logic for stock prediction in TAIFEX. Quantum Evolutionary Algorithm is applied on interval lengths for finding out optimized intervals producing best forecasting accuracy. This algorithm is applied on TAIFEX index prediction and results compared with existing methods. This method is unique in the sense that Quantum computing along with Genetic algorithms and fuzzy logic has never been developed before. The methods provide a new dimension for economic and financial modeling. Contribution 4: Portfolio optimization is a formal approach for financial decision making which holds immense importance for traders, investors and fund managers. For clear and futuristic portfolio management, underlying assets should be optimally classified. A fuzzy granularity based clustering method is proposed for portfolio management, which employs Fuzzy Particle Swarm Optimization to create granules of assets. These granules are then used for portfolio management. This algorithm is applied on listed companies taken from Hong Kong stocks exchange for the period from July 2010 till June 2011. To analyze the performance of proposed algorithm portfolio returns obtained from the proposed method are compared with the portfolio returns of Hong Kong Stock Exchange benchmark index for the duration July 2011 till December 2011. Comparison proved that proposed model’s results are better in comparison to benchmark results of Hang Sang Composite Index. Proposed algorithms are applied on standard benchmark datasets taken from different financial markets. Comparison of results with existing methods proved that the proposed nature inspired methods for economic and financial modeling are better than traditional models in terms of forecasting accuracy and efficiency. The proposed models are robust and can be generalized on other stock markets and assets with minor or no modification. Furthermore the techniques used in this research can be reproduced using various nature inspired methods for stock index prediction and portfolio optimization.