Exploring Hidden Coherent Feature Groups and Temporal Semantics for Multimedia Big Data Analysis

Thanks to the advanced technologies and social networks that allow the data to be widely shared among the Internet, there is an explosion of pervasive multimedia data, generating high demands of multimedia services and applications in various areas for people to easily access and manage multimedia data. Towards such demands, multimedia big data analysis has become an emerging hot topic in both industry and academia, which ranges from basic infrastructure, management, search, and mining to security, privacy, and applications. Within the scope of this dissertation, a multimedia big data analysis framework is proposed for semantic information management and retrieval with a focus on rare event detection in videos. The proposed framework is able to explore hidden semantic feature groups in multimedia data and incorporate temporal semantics, especially for video event detection. First, a hierarchical semantic data representation is presented to alleviate the semantic gap issue, and the Hidden Coherent Feature Group (HCFG) analysis method is proposed to capture the correlation between features and separate the original feature set into semantic groups, seamlessly integrating multimedia data in multiple modalities. Next, an Importance Factor based Temporal Multiple Correspondence Analysis (i.e., IF-TMCA) approach is presented for effective event detection. Specifically, the HCFG algorithm is integrated with the Hierarchical Information Gain Analysis (HIGA) method to generate the Importance Factor (IF) for producing the initial detection results. Then, the TMCA algorithm is proposed to efficiently incorporate temporal semantics for re-ranking and improving the final performance. At last, a sampling-based ensemble learning mechanism is applied to further accommodate the imbalanced datasets. In addition to the multimedia semantic representation and class imbalance problems, lack of organization is another critical issue for multimedia big data analysis. In this framework, an affinity propagation-based summarization method is also proposed to transform the unorganized data into a better structure with clean and well-organized information. The whole framework has been thoroughly evaluated across multiple domains, such as soccer goal event detection and disaster information management.

Marketing practices and their effectiveness in Taiwan's colleges of technology

The main purpose of this study was to investigate marketing practices in Taiwan's institutions of higher education and their relationship with students' college choice behaviors and attitudes. The study was conducted in 11 Taiwan's colleges of technology. It employed a multistrand conversion mixed model design, consisting of a qualitative and a quantitative strand. Funnel-sequenced interviews were conducted with 19 college administrators and the results were content analyzed using a constant-comparative method. The administrator interview data were also quantitized and used in cluster analysis of the institutions. Questionnaire data were collected from 1474 freshmen students, and analyzed using several univariate and multivariate statistical techniques including factor analysis, MANOVA, and correspondence analysis. ^ Analyses indicated that a weak relationship existed between institutions' marketing intensity and students' college choice. Students did not consider institutions' recruitment activities useful in their college searching process. They also reported little knowledge of their current school when they were deciding to enroll. Data analysis also revealed that students were practically oriented in their college selection. Academic resources, employability after graduation, and tuition were the most important attributes in students' college selection. Parents and students' social network such as friends and high school teachers were significant personal sources in enrollment decisions while institutions' representatives (i.e., recruiters) were considered the least influential. ^ Using cluster analysis, institutions were divided into three groups based on intensity of marketing efforts. Multivariate analysis of variance did not reveal significant differences between the college choice behaviors and attitudes of students who entered these three types of institutions. ^ Content analysis of the administrators' interviews indicated that the majority of them practiced passive marketing. This was primarily as a result of resistance to active marketing, lack of leadership commitment, insufficient financial and human resources, little faculty involvement, and inexperience in marketing. In comparison to public institutions, private institutions showed a more favorable attitude towards marketing concepts. They were well advanced in their recruitment activities while public schools were relatively hesitant to use marketing. Curriculum issues were not well represented in marketing activities and did not seem to be impacted by marketing needs. Based on the analysis of qualitative and quantitative data, it can be concluded that in these colleges, curriculum was more driven by commercial and industrial interest than by students' demands. ^ Theoretical, policy, and methodological implementation of the results were discussed. ^

Molecular markers and human history: A tale of two haplogroup systems

To chronicle demographic movement across African Asian corridors, a variety of molecular (sequence analysis, restriction mapping and denaturing high performance liquid chromatography etc.) and statistical (correspondence analysis, AMOVA, calculation of diversity indices and phylogenetic inference, etc.) techniques were employed to assess the phylogeographic patterns of mtDNA control region and Y chromosomal variation among 14 sub-Saharan, North African and Middle Eastern populations. The patterns of genetic diversity revealed evidence of multiple migrations across several African Asian passageways as well within the African continent itself. The two-part analysis uncovered several interesting results which include the following: (1) a north (Egypt and Middle East Asia) to south (sub-Saharan Africa) partitioning of both mtDNA and Y chromosomal haplogroup diversity, (2) a genetic diversity gradient in sub-Saharan Africa from east to west, (3) evidence in favor of the Levantine Corridor over the Horn of Africa as the major genetic conduit since the Last Glacial Maximum, (4) a substantially higher mtDNA versus Y chromosomal sub-Saharan component in the Middle East collections, (5) a higher representation of East versus West African mtDNA haplotypes in the Arabian Peninsula populations versus no such bias in the Levant groups and lastly, (6) genetic remnants of the Bantu demographic expansion in sub-Saharan Africa. ^

Changes in Community Structure of Sediment Bacteria Along the Florida Coastal Everglades Marsh–Mangrove–Seagrass Salinity Gradient

Community structure of sediment bacteria in the Everglades freshwater marsh, fringing mangrove forest, and Florida Bay seagrass meadows were described based on polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) patterns of 16S rRNA gene fragments and by sequencing analysis of DGGE bands. The DGGE patterns were correlated with the environmental variables by means of canonical correspondence analysis. There was no significant trend in the Shannon–Weiner index among the sediment samples along the salinity gradient. However, cluster analysis based on DGGE patterns revealed that the bacterial community structure differed according to sites. Not only were these salinity/vegetation regions distinct but the sediment bacteria communities were consistently different along the gradient from freshwater marsh, mangrove forest, eastern-central Florida Bay, and western Florida Bay. Actinobacteria- and Bacteroidetes/Chlorobi-like DNA sequences were amplified throughout all sampling sites. More Chloroflexi and members of candidate division WS3 were found in freshwater marsh and mangrove forest sites than in seagrass sites. The appearance of candidate division OP8-like DNA sequences in mangrove sites distinguished these communities from those of freshwater marsh. The seagrass sites were characterized by reduced presence of bands belonging to Chloroflexi with increased presence of those bands related to Cyanobacteria, γ-Proteobacteria, Spirochetes, and Planctomycetes. This included the sulfate-reducing bacteria, which are prevalent in marine environments. Clearly, bacterial communities in the sediment were different along the gradient, which can be explained mainly by the differences in salinity and total phosphorus.

The Allelic Abundance of Afghanistan, a Pool for West and South Eurasian Genes and a Blockade against Eastern Admixture

In this study, I divided samples from individuals within Afghanistan based upon geography (i.e., north versus south). I determined allelic frequencies and other statistical parameters for 15 STR loci (i.e., D8S1179, D21S11, D7S820, CSF1PO, D3S1358, TH01, Dl3S317, D16S539, D2S1338, D19S433, vWA, TPOX, D18S51, D5S818, and FGA). I conducted pairwise comparisons with 19 neighboring Eurasian populations to assign Gstatistics and p-values. Categorizing the populations into five groups (i.e., Central Asia, East Asia, South Asia, the Middle East, and the Caucasus/Anatolia), I derived values for intra-population, inter-population, and total variance. Admixture analyses determined the highest allelic contributions to be from the Caucasus/ Anatolia, while negligible contributions were made by Central Asia and East Asia. A Correspondence Analysis revealed clustering of both northern and southern Afghanistan with Georgia, Turkey, northern Iran, and southern Iran of the Caucasus/ Anatolia and the Middle East. A Neighbor-Joining phylogenetic tree was constructed to generate bootstrap values over 1, 000 reiterations.

Use of conjoint analysis to examine beliefs related to the practice of early literacy instruction

The purpose of the study was to examine the relationship between teacher beliefs and actual classroom practice in early literacy instruction. Conjoint analysis was used to measure teachers' beliefs on four early literacy factors—phonological awareness, print awareness, graphophonic awareness, and structural awareness. A collective case study format was then used to measure the correspondence of teachers' beliefs with their actual classroom practice. ^ Ninety Project READS participants were given twelve cards in an orthogonal experimental design describing students that either met or did not meet criteria on the four early literacy factors. Conjoint measurements of whether the student is an efficient reader were taken. These measurements provided relative importance scores for each respondent. Based on the relative important scores, four teachers were chosen to participate in a collective case study. ^ The conjoint results enabled the clustering of teachers into four distinct groups, each aligned with one of the four early literacy factors. K-means cluster analysis of the relative importance measurements showed commonalities among the ninety respondents' beliefs. The collective case study results were mixed. Implications for researchers and practitioners include the use of conjoint analysis in measuring teacher beliefs on the four early literacy factors. Further, the understanding of teacher preferences on these beliefs may assist in the development of curriculum design and therefore increase educational effectiveness. Finally, comparisons between teachers' beliefs on the four early literacy factors and actual instructional practices may facilitate teacher self-reflection thus encouraging positive teacher change. ^