Thesis advisor: Mary E. Walsh / With an increase in emphasis on individual uniqueness and multi-contextual influences, developmental and intervention/prevention science along with similar fields of research (e.g., personalized medicine, personalized learning, health communication, business marketing) have promoted the design and implementation of interventions that would tailor responses and strategies to optimize targeted outcomes based on individual needs and variability (Joyner & Paneth, 2019; Kreuter et al., 1999; Vesanen, 2007). However, in spite of the effort and resources invested in personalization in the past decades, evidence for the realization and utility of tailored interventions have been more anecdotal than quantitatively empirical. The majority of person-centered studies have been qualitative (Lerner et al., 2019). While there is little agreement on what “tailoring” means across the different fields of study, there is a consensus that the term “tailoring” and tailoring-related terms (e.g., personalization, individualization, differentiation, and customization) lack a common and feasible theoretical foundation. Consequently, this semantic crisis has made the construct increasingly difficult to conceptualize and operationalize (e.g. Economist Group, 2021; Shemshack & Spector, 2020). Drawing on insights from the Specificity Principle, Orthogenetic Principle, and Developmental Contextualism in developmental science, this dissertation proposed a provisional definition of “tailoring”: the process of matching unique patterns of services based on each student’s cumulative strengths and needs and the availability of services (e.g., Bornstein, 2015; Lerner et al., 1998; Walsh et al., 2002; Werner & Kaplan, 1956). Guided by this definition, this dissertation sought to find evidence of “tailoring” in one “whole-child,” school-based/evidence-based Integrated Student Support (ISS): City Connects. City Connects partners with school personnel and multiple community agencies to systematically and cost-effectively allocate services/resources to students and their families from low-income communities in order to promote strengths, address needs, and mitigate risks (Moore & Emig, 2014; Dearing et al. 2016; Walsh & Theodorakakis, 2017). After establishing a theoretically-informed basis for “tailoring” as an operationalizable construct, this dissertation employed a comprehensive, three-dimensional approach to data analysis: nomothetic (for finding general/ “universal” trends), differential (for finding differences between groups), and idiographic (for finding differences between individuals) (e.g., Lerner et al., 2019; Overton, 2015; Salvatore & Valsiner, 2010). This was to magnify the descriptive power of the data and findings. In order to accomplish this, the two exploratory substudies in this dissertation employed 1) descriptive analysis, 2) a novel approach for comparing the service patterns matched to each student’s unique sets of strengths and needs, and 3) Latent Class Analysis (LCA). The major findings suggest that “tailoring” in City Connects schools is occurring in five ways: 1) students with higher needs receive more support than students with fewer needs; 2) City Connects is adaptive in responding to the emerging needs of individuals as circumstances change in the course of time; 3) there are unique patterns of services that are either shared (two more students have the same combination of services/types of services) or unshared (only one student has a particular service pattern); 4) service patterns are related to students’ developmental needs (i.e., higher risk level->higher percentages of individualized service patterns); and 5) service pattern matching is purposeful and does not occur randomly. The implications that these findings have on theory, research, and practice are discussed. / Thesis (PhD) — Boston College, 2023. / Submitted to: Boston College. Lynch School of Education. / Discipline: Counseling, Developmental and Educational Psychology.

No / Accurate web log mining results and efficient online navigational pattern prediction are undeniably crucial for tuning up websites and consequently helping in visitors' retention. Like any other data mining task, web log mining starts with data cleaning and preparation and it ends up discovering some hidden knowledge which cannot be extracted using conventional methods. In order for this process to yield good results it has to rely on some good quality input data. Therefore, more focus in this process should be on data cleaning and pre-processing. On the other hand, one of the challenges facing online prediction is scalability. As a result any improvement in the efficiency of online prediction solutions is more than necessary. As a response to the aforementioned concerns we are proposing an enhancement to the web log mining process and to the online navigational pattern prediction. Our contribution contains three different components. First, we are proposing a refined time-out based heuristic for session identification. Second, we are suggesting the usage of a specific density based algorithm for navigational pattern discovery. Finally, a new approach for efficient online prediction is also suggested. The conducted experiments demonstrate the applicability and effectiveness of the proposed approach. (C) 2013 Elsevier B.V. All rights reserved.

No description available.

No description available.

No description available.

No description available.

BACKGROUND: Schizophrenia is a diverse disorder comprising positive, negative, and cognitive symptoms. While several cognitive domains have been alleged to influence symptom severity and quality of life, impaired information processing speed (IPS) is the most profoundly affected and is present even before a diagnosis of a psychotic disorder. Although extensively studied, reliable and reproducible pathological neural circuitry leading to cognitive dysfunction has yet to be fully identified or understood. OBJECTIVE: Our study seeks to identify the neurobiological connections responsible for diminished processing speed in a group of patients with psychotic disorders and a group of individuals at risk to develop a psychotic disorder. METHODS: This data-driven evaluation of resting-state fMRI data and cognitive testing considers two populations: one cohort of adults diagnosed with psychotic disorders and another cohort of individuals identified as being at risk for developing psychotic disorders (ShangHai At-Risk for Psychosis). The images were processed and evaluated using multivariate pattern analysis (MVPA) to compare brain connectivity to information processing speed. The MVPA calculations generated a map to determine the regions where brain connectivity has the strongest correlation to information processing speed as measured by the Brief Assessment in Cognition for Schizophrenia (BACS). RESULTS: The results of the cognitive assessments demonstrate that information processing speed (IPS) is the most impaired domain of cognition in adults with psychotic disorders. Multivariate distance matrix regression (MDMR) was used to identify the cerebellum as the specific brain location where connectivity was most strongly correlated to information processing speed. Specifically, the cerebellar (CBM) connectivity to the dorsal medial prefrontal cortex (dmPFC) exhibited the strongest positive correlation with IPS. The cohort of individuals diagnosed with a psychotic disorder (n = 103) exhibited a moderately strong relationship (r = 0.396, p = 0.000036) between the CBM-dmPFC coactivation and IPS. This finding was then replicated in the SHARP dataset. The clinical high risk sample (n = 137) also identified the CBM-dmPFC circuit as the strongest correlation to information processing speed (r=0.39, p = 0.000002057). Furthermore, the CBM-dmPFC circuit relationship to IPS became stronger as we considered the most impaired subjects. Individuals who would go on to be diagnosed with a psychotic disorder (converters, n = 19) exhibited the strongest relationship between CBM-dmPFC and information processing speed (r = 0.56, p = 0.01). CONCLUSIONS: The reduction in functional connectivity between cerebellum and dmPFC appears to be critically related to the generalized decrease in information processing speed in patients with psychotic disorders and patients at risk of developing a psychotic disorder in the future. The finding is reproduced in two culturally, linguistically, and geographically disparate populations. Interestingly, the finding demonstrates a link between the functional dysconnectivity and cognitive dysfunction in the years before and during the onset of a diagnosis of a psychotic disorder. The research indicates a diagnostic role for assessing information processing speed to identify prodromal patients at risk of developing a psychotic disorder. Finally, the identification of a specific neurological circuit related to impaired cognition would provide a therapeutic target to alleviate cognitive deficits in patients with psychotic disorders in the future.

The waste management and recycling company, Söderhalls Renhållningsverk (SÖRAB) have constructed a surface flow wetland in order to treat surface runoff from the waste management site, Löt. The contaminated water passes several treatment steps until it reaches the wetland and a subsequent soil infiltration step. It is suspected that the flow path of the water through the wetland is short-circuited which may result in a reduced treatment efficiency. The current discharge concentrations of the chemical compounds tested for do not exceed the allowed discharge limits. However, it is of interest to keep the discharge concentrations as low as possible to protect sensitive areas and water bodies downstream. The aim of the thesis was therefore to investigate the flow pattern of the wetland and suggest measures which potentially could improve the treatment efficiency. The flow pattern was modelled numerically in a Physio-Mathematical model developed by Wörman and Kjellin (2020). The current flow pattern was modelled, followed by several simulation runs where the controlling factors of the flow were changed one by one. The validity of the modelling result is uncertain and should therefore be confirmed or rejected by conducting a tracer test prior to implementing any changes in the wetland design. The modelling results indicate the presence of a main flow path passing through the narrower section of the permeable embankment (intersecting the northern and southern part of the wetland, see Fig. 3). The results further indicate that the permeable embankment, the bottom topography and the vegetation distribution were the three major factors controlling the flow pattern within the wetland. Recommended improvements would therefore be to first conduct a tracer test to make sure that any changes implemented are based on the true current flow pattern. The embankment and the vegetation distribution seem to be the main causes of non-idealities in the flow but at the same time probably also have positive effects on the treatment efficiency (since they provide filtration and surface areas where microorganisms can attach to perform their treatment). One solution to reduce the non- idealities in the flow could therefore be to adjust the embankment to be equally wide and deep across the wetland. Furthermore, the bottom topography around the embankment could be adjusted so that the shift in bottom elevation is equal around it. This would probably aid in hindering the development of preferential flow paths through the embankment. Finally, the vegetation distribution could be made uniform. (It should also be noted that a uniform vegetation distribution would require adjustment of the water depth to be below 1,5 m to allow an equal establishment of vegetation). / Avfallshanterings- och återvinningsföretaget Söderhalls Renhållningsverk AB (SÖRAB) har anlagt en våtmark vid Löts avfallsanläggning, ca 35 km norr om Stockholm. Våtmarken utgör det näst sista reningssteget i reningsprocessen av lak- och processvatten från avfallsanläggningen. Riktvärdena för utsläppshalterna överskrids inte men det är önskvärt för SÖRAB att åstadkomma så låga utsläppsvärden som möjligt för att minimera påverkan på känsliga områden och vattendrag nedströms. SÖRAB misstänker dock att preferentiella flödesvägar kan förekomma där vattnet passerar alltför fort genom våtmarken. Detta resulterar ofta i en lägre reningseffektivitet då föroreningarna som är lösta i vattnet får en kortare kontakttid med de naturliga reningsmekanismerna som förekommer i våtmarken. Syftet med det här arbetet var därför att undersöka strömningen i våtmarken och identifiera vattnets flödesvägar med hjälp av en numerisk modell utvecklad av Wörman och Kjellin (2020). Strömningen i våtmarker styrs av en energigradient där flödet går från en punkt i ett vattendrag med högt energiinnehåll till en punkt med lågt energiinnehåll. Enligt principen om energins bevarande omvandlas energin mellan kinetisk, potentiell, tryckenergi och värmeenergi. Ändringen i energiinnehåll beror i sin tur på verkan av externa krafter (gravitationskraften, hydrostatiska tryckkrafter, friktionskrafter, kontraktions- och expansionskrafter och skjuvkrafter från vind. Dessa krafter verkar vid förändringar i bottentopografin, vattendjupet, ytmotståndet (vid våtmarkens botten och väggar), våtmarkens geometri samt där vattenytan är exponerad för vind. Dessa faktorer orsakar utvecklandet av skjuvkrafter i flödet som i sin tur orsakar utvecklandet av hastighetsprofiler och omblandning. Bildandet av hastighetsprofiler och omblandning av vattnet gör att olika vätskeelement eller föroreningar stannar i våtmarken olika lång tid då de rör sig olika fort. För att åstadkomma en så hög reningsgrad som möjligt är det därför önskvärt att hela våtmarkens volym nyttjas till samma grad och att samtliga vätskeelement rör sig med samma hastighet genom våtmarken vilket även kallas för en ideal flödesregim (där ingen omblandning i flödesriktningen förekommer). Då utvecklandet av skjuvkrafter i flödet utgör den grundläggande orsaken till avvikelser från en ideal flödesregim är det önskvärt att minimera dessa. De styrande faktorerna som orsakade utvecklandet av skjuvkrafter i flödet var som nämndes ovan: variationer i våtmarkens bottentopografi och vattendjup, våtmarkens geometri (som kan orsaka isolerade vattenvolymer), ytmotståndet (som bl.a. beror på distributionen av vegetation), vind, samt in- och utflödeshastigeter. Våtmarken undersöktes genom inmätning och lodning som sedan låg till grund för en konceptuell modell av systemet. Den konceptuella modellen utgjorde sedan en grund för att modellera våtmarken numeriskt. I den konceptuella modellen ingick identifiering och definition av systemgränser samt randvillkor, att definiera bottentopografin samt att dela in våtmarken i delområden med homogent flödesmotstånd. Systemgränser och randvillkor identifierades baserat på en vattenbalans. Med hjälp av vattenbalansen kunde relevanta komponenter att inkludera som randvillkor identifieras. Randvillkorens värden bestämdes genom mätningar av den hydrauliska potentialen med hjälp av en GNSS-mottagare (där GNSS står för Global Navigation Satellite System och mottagaren nyttjar satelliter för att bestämma höjd och position i en geografisk punkt) samt från erhållna mätvärden från SÖRAB av in- och utflöden som pumpas in och ut ur våtmarken. Då bottentopografin var okänd bestämdes den genom att vattendjupet mättes, interpolerades och subtraherades från en referensnivå. Vattendjupet mättes genom lodning och mätpunkterna interpolerades sedan i det geografiska informationsverktyget ArcMap för att erhålla ett heltäckande lager av mätvärden över vattendjupet. Därefter kunde vattendjupet subtraheras från vattenytans medelhöjd över havet vilket var 38,1 m och ett heltäckande lager över bottentopografin kunde erhållas. Flödesmotståndet beräknades enligt ekvation 16 (se avsnitt 2.1.) och våtmarken delades in i homogena delområden baserat på förekomsten av vegetation och om strömningen skedde genom ett poröst medium eller öppet vatten. Den numeriska modelleringen genomfördes genom att först modellera den nuvarande strömningen med och utan ett rör som går genom den genomsläppliga vallen. Därefter ändrades de styrande faktorerna för flödet en i taget för att kunna utvärdera vilken effekt varje faktor hade på strömningen i våtmarken för att åstadkomma en ideal flödesregim. Följande simuleringar genomfördes: (1) Bottentopografin gjordes jämn med ett vattendjup på 0,5 respektive 1,0 m, (2) inloppszonen gjordes längre för att undvika skapandet av isolerade vattenvolymer längs ytterkanterna, (3) vegetationens distribution gjordes homogen, (4) den genomsläppliga vallen justerades. Pålitligheten i resultatet från den numeriska modelleringen är osäker. Därför skulle ett spårämnesförsök behöva genomföras för att undersöka huruvida resultatet från modelleringen är tillförlitligt eller inte. Enligt modelleringsresultatet förekom en tydlig flödesväg som passerade genom den smalare delen av vallen. Modelleringen indikerade vidare att de kontrollerande faktorerna som styrde flödet i våtmarken framförallt utgjordes av distributionen av vegetation samt den genomsläppliga vallen tillsammans med omgivande bottentopografi. Det rekommenderas därför att modelleringsresultatets validitet först undersöks med ett spårämnesförsök innan några åtgärder vidtas. Möjliga förbättringsåtgärder som skulle kunna införas därefter för att sträva mot ett idealt flöde skulle kunna vara att justera den genomsläppliga vallen till att ha en lika stor tjocklek och ett lika stort djup överallt samt att justera bottentopografin runt vallen så att den förändras lika mycket runt vallen. Detta för att skapa förutsättningar för att undvika preferentiella flödesvägar genom den. En jämn fördelning av vegetation (och justering av vattendjupet till att understiga 1,5 m som tillåter kolonisering av växter) att störningar i flödet minimeras.

A regional IPM project was initiated with four rounds of sampling for white grubs on the fairways of nine golf courses located on the Lower Peninsula of eastern Virginia, from 2000 through 2002. Fifteen regressor variables were collected and measured that included local-scale variables, golf course management practices and spatial pattern metrics derived from satellite images that underwent two methods of a supervised classification of six land-cover types (turf, woods, wetland, urban, bare soil and water) on four landscape scales derived from 10 km x 10 km buffer zones surrounding each golf course. Pearson's correlation coefficients were calculated to reduce the number of variables to a few that were highly correlated with white grub densities. Mallow's C(p) calculations were performed on the reduced variable sets to extract those that would be highly predictive. A multiple linear regression was performed using the Mallow's variables to develop eight regression equations (two classification methods x four landscape scales) that were used to predict regional white grub presence and abundance in 2003 on six additional golf courses located on the Lower Peninsula. The best model was the 6 km x 6 km buffer zones model from the second classification method, which included one local-scale variable (golf course age) and three spatial pattern metrics (total turf area, total turf area-to-total urban area ratio, and a woods interspersion-juxtaposition index). The mean difference between actual and predicted values was -0.15, standard deviation = 0.79, R2 = 81.38%. Additionally, a study was conducted to determine whether the number of white grubs collected from transects of sampled golf course fairways was significantly different from those found in the roughs. White grub counts from the roughs were significantly higher (mean = 0.283 grubs/transect, standard error = 0.0135) than those from fairways (mean = 0.146 grubs/transect, standard error = 0.0188); t = -4.31, df = 735, P = 0.0001. / Ph. D.

Event-related collections, including both tweets and webpages, have valuable information, and are worth exploring in interdisciplinary research and education. Unfortunately, such data is noisy, so this variety of information has not been adequately exploited. Further, for better understanding, more knowledge hidden behind events needs to be unearthed. Regarding these collections, different societies may have different requirements in particular scenarios. Some may need relatively clean datasets for data exploration and data mining. Social researchers require preprocessing of information, so they can conduct analyses. General societies are interested in the overall descriptions of events. However, few systems, tools, or methods exist to support the flexible use of event-related collections. In this research, we propose a new, integrated system to process and analyze event-related collections at different levels (i.e., data, information, and knowledge). It also provides various services and covers the most important stages in a system pipeline, including collection development, curation, analysis, integration, and visualization. Firstly, we propose a query likelihood model with pre-query design and post-query expansion to rank a webpage corpus by query generation probability, and retrieve relevant webpages from event-related tweet collections. We further preserve webpage data into WARC files and enrich original tweets with webpages in JSON format. As an application of data management, we conduct an empirical study of the embedded URLs in tweets based on collection development and data curation techniques. Secondly, we develop TwiRole, an integrated model for 3-way user classification on Twitter, which detects brand-related, female-related, and male-related tweeters through multiple features with both machine learning (i.e., random forest classifier) and deep learning (i.e., an 18-layer ResNet) techniques. As guidance to user-centered social research at the information level, we combine TwiRole with a pre-trained recurrent neural network-based emotion detection model, and carry out tweeting pattern analyses on disaster-related collections. Finally, we propose a tweet-guided multi-document summarization (TMDS) model, which generates summaries of the event-related collections by using tweets associated with those events. The TMDS model also considers three aspects of named entities (i.e., importance, relatedness, and diversity) as well as topics, to score sentences in webpages, and then rank selected relevant sentences in proper order for summarization. The entire system is realized using many technologies, such as collection development, natural language processing, machine learning, and deep learning. For each part, comprehensive evaluations are carried out, that confirm the effectiveness and accuracy of our proposed approaches. Regarding broader impact, the outcomes proposed in our study can be easily adopted or extended for further event analyses and service development. / Doctor of Philosophy / Event-related collections, including both tweets and webpages, have valuable information. They are worth exploring in interdisciplinary research and education. Unfortunately, such data is noisy. Many tweets and webpages are not relevant to the events. This leads to difficulties during data analysis of the datasets, as well as explanation of the results. Further, for better understanding, more knowledge hidden behind events needs to be unearthed. Regarding these collections, different groups of people may have different requirements. Some may need relatively clean datasets for data exploration. Some require preprocessing of information, so they can conduct analyses, e.g., based on tweeter type or content topic. General societies are interested in the overall descriptions of events. However, few systems, tools, or methods exist to support the flexible use of event-related collections. Accordingly, we describe our new framework and integrated system to process and analyze event-related collections. It provides varied services and covers the most important stages in a system pipeline. It has sub-systems to clean, manage, analyze, integrate, and visualize event-related collections. It takes an event-related tweet collection as input and generates an event-related webpage corpus by leveraging Wikipedia and the URLs embedded in tweets. It also combines and enriches original tweets with webpages. As an application of data management, we conduct an empirical study of tweets and their embedded URLs. We developed TwiRole for 3-way user classification on Twitter. It detects brand-related, female-related, and male-related tweeters through their profiles, tweets, and images. To aid user-centered social research, we combine TwiRole with an existing emotion detection tool, and carry out tweeting pattern analyses on disaster-related collections. Finally, we propose a tweet-guided multi-document summarization (TMDS) model and service, which generates summaries of the event-related collections by using tweets associated with those events. It extracts important sentences across different topics from webpages, and organizes them in proper order. The entire system is realized using many technologies, such as collection development, natural language processing, machine learning, and deep learning. For each part, comprehensive evaluations help confirm the effectiveness and accuracy of our proposed approaches. Regarding broader impact, our methods and system can be easily adopted or extended for further event analyses and service development.