Browsing by Subject "models"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Community-Oriented Models and Applications for the Social Web(2012-07-16) Kashoob, Said Masoud AliThe past few years have seen the rapid rise of all things "social" on the web from the growth of online social networks like Facebook, to user-contributed content sites like Flickr and YouTube, to social bookmarking services like Delicious, among many others. Whereas traditional approaches to organizing and accessing the web?s massive amount of information have focused on content-based and link-based approaches, these social systems offer rich opportunities for user-based and community-based exploration and analysis of the web by building on the unprecedented access to the interests and perspectives of millions of users. We focus here on the challenge of modeling and mining social bookmarking systems, in which resources are enriched by large-scale socially generated metadata (?tags?) and contextualized by the user communities that are associated with the resources. Our hypothesis is that an underlying social collective intelligence is embedded in the uncoordinated actions of users on social bookmarking services, and that this social collective intelligence can be leveraged for enhanced web-based information discovery and knowledge sharing. Concretely, we posit the existence of underlying implicit communities in these social bookmarking systems that drive the social bookmarking process which can provide a foundation for community-based organization of web resources. To that end, we make three contributions: ? First, we propose a pair of novel probabilistic generative models for describing and modeling community-oriented social bookmarking. We show how these models enable effective extraction of meaningful communities over large real world social bookmarking services. ? Second, we develop two frameworks for community-based web information browsing and search that are based on these community-oriented social bookmarking models. We show how both achieve improved discovery and exploration of the social web. ? Third, we introduce a community evolution framework for studying and analyzing social bookmarking communities over time. We explore the temporal dimension of social bookmarking and explore the dynamics of community formation, evolution, and dissolution. By uncovering implicit communities, putting them to use in an application scenario (search and browsing), and analyzing them over time, this dissertation provides a foundation for the study of how social knowledge networks are self-organized, a deeper understanding and appreciation of the factors impacting collective intelligence, and the creation of new information access algorithms for leveraging these communities.Item Fundamental study of structural features affecting enzymatic hydrolysis of lignocellulosic biomass(Texas A&M University, 2006-10-30) Zhu, LiLignocellulose is a promising and valuable alternative energy source. Native lignocellulosic biomass has limited accessibility to cellulase enzyme due to structural features; therefore, pretreatment is an essential prerequisite to make biomass accessible and reactive by altering its structural features. The effects of substrate concentration, addition of cellobiase, enzyme loading, and structural features on biomass digestibility were explored. The addition of supplemental cellobiase to the enzyme complex greatly increased the initial rate and ultimate extent of biomass hydrolysis by converting the strong inhibitor, cellobiose, to glucose. A low substrate concentration (10 g/L) was employed to prevent end-product inhibition by cellobiose and glucose. The rate and extent of biomass hydrolysis significantly depend on enzyme loading and structural features resulting from pretreatment, thus the hydrolysis and pretreatment processes are intimately coupled because of structural features. Model lignocelluloses with various structural features were hydrolyzed with a variety of cellulase loadings for 1, 6, and 72 h. Glucan, xylan, and total sugar conversions at 1, 6, and 72 h were linearly proportional to the logarithm of cellulase loadings from approximately 10% to 90% conversion, indicating that the simplified HCH-1 model is valid for predicting lignocellulose digestibility. Carbohydrate conversions at a given time versus the natural logarithm of cellulase loadings were plotted to obtain the slopes and intercepts which were correlated to structural features (lignin content, acetyl content, cellulose crystallinity, and carbohydrate content) by both parametric and nonparametric regression models. The predictive ability of the models was evaluated by a variety of biomass (corn stover, bagasse, and rice straw) treated with lime, dilute acid, ammonia fiber explosion (AFEX), and aqueous ammonia. The measured slopes, intercepts, and carbohydrate conversions at 1, 6, and 72 h were compared to the values predicted by the parametric and nonparametric models. The smaller mean square error (MSE) in the parametric models indicates more satisfactorily predictive ability than the nonparametric models. The agreement between the measured and predicted values shows that lignin content, acetyl content, and cellulose crystallinity are key factors that determine biomass digestibility, and that biomass digestibility can be predicted over a wide range of cellulase loadings using the simplified HCH-1 model.Item Seal strength models for medical device trays(2009-05-15) Mays, Patricia FayeSeven empirical equations were developed for the prediction of seal strength for medical device trays. A new methodology was developed and used for identifying burst and peel locations and comparing burst pressure and peel force. Multiple linear regression was used to fit 76 models, selecting the best models based on the Akaike Information Criterion (AIC) and adjusted R2 (R2 adj) value of each model. The selected models have R2 adj and prediction R2 (R2 pred) values of .83 to .94. Factors investigated for the peel force response were sealing pressure (3 levels), dwell time (3 levels), sealing temperature (3 levels), and adhesive. Additional factors investigated for the burst pressure response were restraining plate gap, and tray volume, height, length-to-width ratio and area. Polyethylene terephthalate-glycol (PETG) trays with Tyvek 1073B lids and two popular water-based adhesives were used. Trays were selected to yield three levels of area and three levels of length-to-width ratio, defining nine package configurations. Packages for burst testing were sealed under a fractional factorial design with 27 treatments. Packages for peel testing were sealed under a 17-point face-centered central composite design. Packages were tested using peel testing following the ASTM F88-07 standard and restrained burst testing with three gap distances following the ASTM F2054-00 standard. All possible subsets of the factors were evaluated, with the best models selected based on AIC value. Equations were developed to predict peak and average peel force based on sealing process parameters (R2 pred =.94 and .92), burst pressure based on tray and sealing parameters and gap (R2 pred =.94), and four peel force responses based on burst pressure and gap (R2 pred =.83 to .86). Models were validated through cross-validation, using the prediction error sum of squares (PRESS) statistic. The R2 pred was calculated to estimate the predictive ability of each model.Item Technology Characterization Models and Their Use in Designing Complex Systems(2011-08-08) Parker, Robert ReedWhen systems designers are making decisions about which components or technologies to select for a design, they often use experience or intuition to select one technology over another. Additionally, developers of new technologies rarely provide more information about their inventions than discrete data points attained in testing, usually in a laboratory. This makes it difficult for system designers to select newer technologies in favor of proven ones. They lack the knowledge about these new technologies to consider them equally with existing technologies. Prior research suggests that set-based design representations can be useful for facilitating collaboration among engineers in a design project, both within and across organizational boundaries. However, existing set-based methods are limited in terms of how the sets are constructed and in terms of the representational capability of the sets. The goal of this research is to introduce and demonstrate new, more general set-based design methods that are effective for characterizing and comparing competing technologies in a utility-based decision framework. To demonstrate the new methods and compare their relative strengths and weaknesses, different technologies for a power plant condenser are compared. The capabilities of different condenser technologies are characterized in terms of sets defined over the space of common condenser attributes (cross sectional area, heat exchange effectiveness, pressure drop, etc.). It is shown that systems designers can use the resulting sets to explore the space of possible condenser designs quickly and effectively. It is expected that this technique will be a useful tool for system designers to evaluate new technologies and compare them to existing ones, while also encouraging the use of new technologies by providing a more accurate representation of their capabilities. I compare four representational methods by measuring the solution accuracy (compared to a more comprehensive optimization procedure's solution), computation time, and scalability (how a model changes with different data sizes). My results demonstrate that a support vector domain description-based method provides the best combination of these traits for this example. When combined with recent research on reducing its computation time, this method becomes even more favorable.