With the development of information technology¡Athe interface of communication is changed into the digit form of the electronic file from the traditional way gradually∙We want make a Chinese font of script that has the characteristics of user¡Alet the interface they used become a script interface with characteristics of other users¡Awhen they are communicate with each other∙ The quantity of the Chinese character is very huge¡Ato deal with all Chinese characters step by step is almost impossible∙In order to solve this problem¡Awe need to reduce the problem level from Chinese character to components of Chinese characters¡Abecause the components of Chinese characters are used as radical in many characters¡Aby this way¡Awe can use fewer input Chinese characters to make a large amount of Chinese characters personalizing∙ First¡Aanalyse all characters in True Type Font file∙Every character has a group of control points∙Calculate the spatial relationship between each control point and other adjacent control points¡Ato judge what stroke type that control point have¡Athen analyse all kinds of stroke types in character to check if there is the characteristics of control points can make some components of Chinese characters∙ Second¡Acalculate features of handwritten radicals in input handwritten Chinese characters which radicals we can recognize in first step¡Athen modify all control points in radical by this features to make radical personalizing∙ Final¡Areplace the personalized radical into all characters which has this radical as components of character to get a personalized True Typr Font file∙

With the advent of modern wireless communication techniques, mobile devices play an increasingly important role in our daily life. Users often use these mobile devices to record and retrieve information about tasks and data. In fact, many of people¡¦s daily activities are not independent, and they are likely to be process-oriented. Previous work reported in [Chen01] used metagraph to model user¡¦s activities and proposed a client-server architecture of personal workflow system for supporting pervasive computing. This architecture allowed users to define their personal processes and provided a set of operations for manipulating them. However, the proposed system is passive in that it did not provide any support to help mobile users construct their personal processes, nor did it provide friendly query interface for users. Our work aims at constructing a personalized workflow template provider and providing a user-friendly interface design. The personalized workflow template provider intends to provide personalized workflow templates of organization-driven processes. Users with different backgrounds and interests may retrieve different personal workflow template to accomplishing the same goal. In client side, we provide five pre-defined queries for mobile user can efficiently inquire a personal process and design a better interface for a mobile user to browse his/her personal processes. At last, we implemented these two components and evaluate the effectiveness of the proposed design strategies.

Given the emerging importance of personalized medicine (PM) in primary care, now should be the ideal time for engaging with health care professionals (HCPs), both physicians and nurses, about integrating PM into practice. The question then becomes: what is the most effective way to engage with HCPs about emerging technologies that are not in routine clinical use and which are unfamiliar to many? The overall aim of this pilot study was to develop and compare two professional engagement (PE) approaches for engaging with HCPs about PM to inform their development and design of a future formal evaluation. The first PE intervention was a structured in-person focus group and the second was an online version, also incorporating an educational component, but without group interaction. The pilot study showed that while participants evaluated both interventions positively, the in-person workshop consistently scored higher; however, recruitment challenges were a major obstacle for this approach.

There has been increasing development in personalized interventions that are tailored to uniquely evolving health status of each patient over time. In this dissertation, we investigate two problems: (1) the construction of individualized mobile health (mHealth) application recommender system; and (2) the estimation of optimal dynamic treatment regimes (DTRs) from a multi-stage clinical trial study. The dissertation is organized as follows. In Chapter 1, we provide a brief background on personalized medicine and two motivating examples which illustrate the needs and benefits of individualized treatment policies. We then introduce reinforcement learning and various methods to obtain the optimal DTRs as well as Q-learning procedure which is a popular method in the DTR literature. In Chapter 2, we propose a partial regularization via orthogonality using the adaptive Lasso (PRO-aLasso) to estimate the optimal policy which maximizes the expected utility in the mHealth setting. We also derive the convergence rate of the expected outcome of the estimated policy to that of the true optimal policy. The PRO-aLasso estimators are shown to enjoy the same oracle properties as the adaptive Lasso. Simulations and real data application demonstrate that the PRO-aLasso yields simple, more stable policies with better results as compared to the adaptive Lasso and other competing methods. In Chapter 3, we propose a penalized A-learning with a Lasso-type penalty for the construction of optimal DTR and derive generalization error bounds of the estimated DTR. We first examine the relationship between value and the Q-functions, and then we provide a finite sample upper bound on the difference in values between the optimal DTR and the estimated DTR. In practice, we implement a multi-stage PRO-aLasso algorithm to obtain the optimal DTR. Simulation results show advantages of the proposed methods over some existing alternatives. The proposed approach is also demonstrated with the data from a depression clinical trial study. In Chapter 4, we present future work and concluding remarks.

Master of Science / Department of Computer Science / William H. Hsu / This report presents a personalized exercise training chatbot for individual users based on data collected from the Internet of Things (IoT), particularly wearable fitness devices. The chatbot is designed with our goal of motivating users to exercise more by discussing exercise statistics with the user, such as whether their daily steps have increased, decreased, or remained steady. In this work I first survey a few examples of how increased interest in fitness and the promotion of healthy lifestyles is driving demand for personalized artificial intelligence, wear- able computing, and ubiquitous computing applications. Next, I describe the design of a data-driven ”personal trainer” chatbot. I then develop a prototype persuasion system based on interactive dialogs delivered via a front-end application, that collects data from wearable equipment using back-end data loggers that I instrumented as a mobile application. Finally, I describe the process of deploying and demonstrating this prototype along with technical challenges and early findings. The overall system consists of (1) the back-end Coach agent, an Android application that collects data from all wearable instruments, and (2) the front-end Me agent, which initiates and continues conversations with the user using notifications that are in turn based on data from the Coach agent. This data-driven ensemble reminds the user to exercise and also gives the user a chance to provide feedback via human/agent interactive dialogs. In this project, I used only one wearable device, the MI Band 2, and get real-time steps and weekly step aggregates from it. The human/agent dialogues are deployed via the Slack groupware platform. Google Sheets is used as a web service for updating and exchanging data.

Though racial classifications may serve as a mechanism for identifying and correcting disparities among various groups, using such classifications in a clinical setting to detect and treat patient needs can be problematic. This report explores how medical professionals and researchers use race in health care for purposes of data collection, risk assessment, and diagnosis and treatment options. Using mixed race individuals as an example, it then discusses some of the problems associated with using race to group individuals, assess risk, and inform patient care. Finally, it discusses how certain components of personalized medicine, such as genetic testing, Electronic Health Records, and Rapid Learning Systems could help address some of the concerns that arise from the application of race in a health care setting. / text

There have been many scientific developments in the last century including the atomic bomb and DNA sequencing. Moreover, when human genome was sequenced in the early 2000s, it opened a new avenue to study disease and human development. Genetic tests have become an integral part for cancer diagnosis. Still, cancer therapy is decided based on the tumor genotype, the very definition of pharmacogenetic testing. More specifically, pharmacogenetics or pharmacogenomics is defined as variations in genes that can affect drug response. There has been great deal of research into pharmacogenetics and its potential fields for application. One such field is cardiology and cardiovascular disease. There are some promising researches that indicate genetic influence over drug response, such as the role of CYP2C19 over metabolism of a drug used for treating acute coronary disease and other cardiovascular issues. This is a great tool in the transition toward personalized medicine; however there are some logistical and social concerns over genetic tests; test administration, result accuracy and validity, data storage and security. Also, many patients were concerned with confidentiality, payment method and timely intervention. Also, implementation plans should include all areas, not just cities. Although there is potential for pharmacogenetic testing, many challenges have to be considered and addressed to ensure public confidence and proper use of the technique. Pharmacogenetics is a step towards individualized or personalized medicine; in-depth research prior to implementation will help tackle any challenges that may arise.

Modern Electronic Health Record (EHR) systems are now integral to healthcare. Having evolved from hospital billing and laboratory systems in the 80's, EHR systems have grown considerably as we learn to represent more and more aspects of patient encounter, diagnosis and treatment digitally. EHR user interfaces, however, lag considerably behind their consumer-electronics counterparts in usability, most notably with respect to customizability. This limitation is especially evident in the implementation of audible alerts that are coupled to sensors or timing devices in intensive-care settings. The most current standard, (ISO/IEC 60601-1-8) has been designed for alerts that are intended to signal situations of varying priorities: however, it is not universally implemented, and has been criticized for the difficulty that healthcare providers have in discriminating between individual alarms, and for the failure to incorporate prior research with respect to "sense of urgency" as it applies to alarm efficacy. In the present work, however, we consider that there are more effective means to allow a user to identify an alarm correctly than "sense of urgency" response. This thesis focuses on the problem of correct identification of alerts: what happens when a human subject is allowed to create or designate (i.e., personalize) one's own alerts? Given the ubiquity, low costs and commoditization of consumer-electronics devices, we believe that it is just a matter of time before such devices become the norm in critical care and replace existing, special-purpose devices for information delivery at the point of patient care. We built a tool, PASA (Personalized Alert Study Application), that would allow us to capture and edit sounds and orchestrate studies that would contrast any two types of sounds. PASA facilitated a study where study participant's responses to "personalized" sounds were contrasted with sounds that meet the ISO/IEC 60601-1-8:2012 standard. We performed two sub-studies that contrasted responses to two banks of 6-alerts and 10-alerts. The 6-alert study was repeated with the same subjects after two weeks without training to measure recall. We observed that accuracy, reaction time, and retention were significantly improved with the personalized sounds. For example, the median errors for the 6-alert baseline study were 4 for personalized vs. 27 for standard alerts. For the 6-alert repeat study it was 7 vs. 43. The median for the 10-alert study was 1 for personalized vs. 55 for standard alerts. Accuracy for recognition, while remaining constant for personalized alerts, degraded considerably for standardized alerts as the number of alerts increased from 6 to 10. We conclude that personalization of alerts may improve information delivery and reduce cognitive overload on the health care provider. This potential positive effect at the point of patient care merits further studies in a clinical or simulated clinical setting.

Cancer therapy has historically been determined by a tumor’s tissue of origin. Now, thanks to advances in genomics technology, scientists are looking further into one’s cancer; into the very genome that drives the tumor growth. The growth of genomics in cancer research has been astronomical. In a little over ten years since the completion of the Human Genome Project, genomic profiling technologies have developed into an incredibly powerful, relatively cheap, and immensely underutilized tool for oncologists. In the midst of the advances in cancer profiling, there has been reluctance from oncologists to incorporate genomic profiling into their treatment decisions. Saddled by outdated clinical trial designs, and cancer drug regulation programs, the true measure of the clinical utility of genomic profiling has yet to be seen. Cancer scientists will continue to profile cancers at a pace well beyond the limits of the field of oncology. Without coordinated efforts to update the oncology healthcare system, compendia of data will continue to be generated with limited ability to translate the information into personalized medicines. There are significant barriers to overcome before genomic data can universally be incorporated into the daily practice of cancer medicine. In the meantime, resources are available for physicians to help begin the process of integrating a more personalized approach to cancer therapy. Third-party bioinformatics companies are in the best position to be the agents of this change. As cancer research continues to adopt a genomic approach, it is paramount that, for the sake of millions of cancer patients, the healthcare system adapts in a way to best utilize this new information.

The research of my dissertation studies the methods of designing and analyzing sequential multiple assignment randomized trial (SMART) for comparing multiple adaptive interventions. As a SMART typically consists of numerous adaptive interventions, inferential procedures based on pairwise comparisons of all interventions may suffer substantial loss in power after accounting for multiplicity. I address this problem using two approaches. First, I propose a likelihood-based Wald test, study the asymptotic distribution of its test statistics, and apply it as a gate-keeping test for making an adaptive intervention selection. Second, I consider a multiple comparison with the best approach by constructing simultaneous confidence intervals that compare the interventions of interest with the truly best intervention, which is assumed to be unknown in inference; an adaptive intervention with the proposed interval excluding zero will be declared as inferior to the truly best with a pre-specified confidence level. Simulation studies show that both methods outperform the corresponding multiple comparison procedures based on Bonferroni's correction in terms of the power of test and the average width of confidence intervals for estimation. Simulations also suggest desirable properties of the proposed methods. I apply these methods to analyze two real data sets. As part of the dissertation, I also develop a user-friendly R software package that covers many statistical work related to SMART, including study design, data analysis and visualization. Both proposed methods can be implemented by using this R package. In the end of the dissertation, I show an application of designing a SMART to compare multiple patient care strategies for depression management based on one of the proposed methods.