tokyo-header-2019.jpg

Tokyo 2019

November 15

Abstracts

vic-strecher-keynote.jpg

Ikigai or Life On Purpose: How Living for What Matters Most Changes Everything

Vic Strecher, Life On Purpose: How Living for What Matters Most Changes Everything

In Japan, it is called “Ikigai” – a reason for being. In the West, it is called “purpose in life.” Both terms relate to applying our best selves to our mission, passion, profession, and vocation. In this presentation, Dr. Strecher uses a broad spectrum of perspectives, from philosophy to neuroscience, to find and follow our Ikigai.

john-sall-keynote.jpg

Import. Understand. Communicate.

John Sall, SAS

JMP 15 and JMP Pro 15 give you more options for exploring and analyzing your data interactively. From enhanced data preparation and modeling to instant graphical gratification, you’ll discover new ways to access and understand your data. You’ll also find a new way to publish, share and communicate with our latest product, JMP Live.

JMP Live: Collaborative analytics software for sharing interactive reports

Jon Weisz, Senior Vice President, JMP Sales and Marketing, SAS Institute Inc.

jon-weisz-keynote

Daniel Valente, Manager, JMP Product Management, JMP WW Marketing, SAS Institute Inc.

dan-valente

Eric Hill, Distinguished Software Developer, JMP Development, SAS Institute Inc.

eric-hill

JMP® Live is new product added to the JMP Family with the release of JMP 15. It is collaborative analytics software for publishing active reports to a private server, supporting filters and column switchers. JMP continues to be a desktop product, but now it can publish live reports, accessible to everyone in the private community, not just JMP users. This talk will cover basic JMP live architectures and deployment, including authentication and report authorization. It will also demonstrate both automated and ad hoc publishing features from JMP. With JMP live, publishing, sharing and community become new strengths.

  • Session ID: 2019-JP-25MP-B4

    Room B

    Among cases of out-of-hospital cardiopulmonary arrest that occurred in a residence, did oral instruction for chest compression by dispatchers improved the initial waveform into shockable forms that could be resuscitated with an AED by adjusting the time between recognition and contact, age, and sex? Cases aged 15-84 years witnessed and reported by citizens with brain function prognosis CPC 1-5 after a month were researched among the Utstein Style data in 2015 to extract 19,062 cases where the event occurred in the order of witness, recognition, and contact and time between recognition and contact was within the 99th percentile. Then, the hazard ratio for the presence of a shockable rhythm was obtained by the Cox proportional hazard model adjusted for age and 119-contact time from 15,500 cases capable of being linked to the CPA occurrence site in the emergency transfer data. Among the 15,500 cases, 9,091 cases (58.6%) occurred in a residence where the bystander was family, excluding cases with defibrillation by a citizen. The hazard ratio for CPR under oral instruction compared to CPR without instruction was 1.44 (1.06-1.97) for 15-64 years of age, 1.08 (0.78-1.54) for 65-74 years of age, and 0.76 (0.56-1.05) for 75-84 years of age. Among cases of out-of-hospital cardiopulmonary arrest that occurred in a residence and were attended by family, the shockable rhythm increased by oral instruction by a dispatcher compared to those with CPR without instruction in the group of 15-64 years of age. From now on, proactive oral instruction to family leads to increases in the number of cardiopulmonary arrest patients who return to society. In the presentation, the analysis results are reported taking into consideration newly published data from 2016 and thereafter.

  • Session ID: 2019-JP-25MP-C3

    Room C

    JMP Clinical is a tool to analyze clinical study data easily, and CDISC standard data should be prepared when using JMP Clinical. If CDISC standard data can be prepared while the clinical study is being performed, the statistical analysis function, one of the features of JMP Clinical, can identify a medical institution with a different trend in the data. We prepared CDISC standard data while we were performing a clinical study and experienced central monitoring (data review) using JMP Clinical. To perform central monitoring, a protocol should be prepared to describe the procedure for a data review. Thus, what to describe in the Central Monitoring Protocol when using JMP Clinical was studied. This presentation introduces the results of this study.

  • Session ID: 2019-JP-50MP-C1

    Room C

    Comprehensive design means the items necessary for design put together. These include comprehensive design in the narrow sense (consisting of three stages: identification, modeling and optimization) and comprehensive design in the broad sense (consisting of five stages: those listed above and feasibility assessment and regression adjustment). A good approximate model is essential in the scientific design. This acquisition requires the selection of factors and the identification of construction (judgment of the presence of product term and quadratic term), and this is the stage of identification. This study performs the identification by adding the identification of construction to the screening for selecting factors. For this, the devised experiment (screening plan added with fixed points) and the statistical processing thereafter are decisive. Determination of a model by estimating the coefficient of the necessary term based on the identification result is the stage of modeling. Here the utilization of a custom plan and an expansion plan is decisive. The last is the stage of optimization for determining the preferable level that achieves the design purpose. Here multipurpose optimization is decisive. By the way, since the solution of the optimization is a hypothesis, the implementation of feasibility assessment is important. Further, if the solution is unacceptable, the adjustment using the regression is performed. The addition of these two stages makes the design in the business certain.

  • Session ID: 2019-JP-25MP-C2

    Room C

    Effective Communication and Visualization With Life Sciences Data

    Kelci Miclaus, Senior Manager Advanced Analytics R&D, JMP Life Sciences, SAS

    Conducting translational and clinical research comes with a big price tag. Given costs and time considerations in studies for understanding health outcomes and disease, researchers commonly collect as much data on as many endpoints as possible. This is further motivated by our limited understanding of the genomic underpinnings of disease and good clinical research practice protocols to assess not only efficacy of new therapeutics, but safety and operational integrity In the life science data “life cycle,” the data volume poses challenges to analyze and communicate results; traditional practices that produce hundreds of tables are neither efficient nor effective. This presentation will focus on visualization to communicate results in several case study analyses typical of life science research. We highlight both data summary visualization techniques and graphs to communicate the results of complex statistical analyses. Examples include volcano plots and Manhattan plots when performing thousands to millions of statistical models with genomic data, distributional summaries of safety and efficacy in clinical trials, laboratory findings trends (waterfall plots, shift plots, spaghetti plots, lasagna plots, swimmer plots) and clinical operational integrity anomaly detection. All examples will be presented with Graph Builder, showcasing the strength and flexibility of this key JMP platform.

  • Session ID: 2019-JP-25MP-C4

    Room C

    Element Analysis for Work Standardization in Video Manual Preparation

    Toshiharu Taguchi, Assistant Manager, ThreeBond Fine Chemical Co., Ltd.

    We are promoting the use of video manuals as teaching materials for the education of new members and overseas staff. However, an issue was found during the manual preparation that non-uniform actions among workers affect the inspection value. When asking a worker “Why do you take that action?” the person could not specifically explain because that action was taken unconsciously. Therefore, in order to perform standardization of work, we first photographed the work condition of each worker and performed the decomposition of the actions into elements. We analyzed the results of the decomposition using JMP and evaluated by an investigation of which action affected the inspection value.

  • Session ID: 2019-JP-50MP-A1

    Room A

    Pictures From the Gallery 4: Select Advanced Graph Builder Views

    Scott Wise, Senior Manager, Analytical Education, SAS

    If a picture is worth a thousand words, then the visuals that can be created in JMP Graph Builder could be considered fine art works with their powerful ability to convey compelling information to the viewer. This journal presentation features how to build popular and captivating advanced graph views using JMP Graph Builder. Based on the popular Pictures From the Gallery journal format, we have created a brand-new Pictures From the Gallery 4 journal that features additional views available in the latest versions of JMP. We will feature several popular industry graph formats that you may not have known could be easily built within JMP. Views such as incorporating butterfly plots, advanced Pareto plots, Gantt charts, graphs in maps and more will be included that can help breathe life into your analytics and provide views that help manage up your results.

  • Session ID: 2019-JP-25MP-B3

    Room B

    One of the operations of clinical engineers is the maintenance and inspections of portable medical devices that operate in the hospital. Although it is desirable to use medical devices within the period of endurance described in the Package Insert or User’s Manual, medical devices may be used for a longer period with repeated repairs. Here we conducted an analysis of 36 small infusion pumps. Our hospital uses the pumps for continuous administration of narcotics to alleviate pain. The devices are often carried by patients who use the device and not controlled only by medical professionals. They often exhibit errors from dropping or strong impacts. The device is designed to store operation logs, and we investigate in response to every request for repairs and inspections due to abnormalities. This investigation has been a burden because it takes time. It is also known empirically that repair costs are expensive, and repairs are frequent. We believe that consideration is needed not only in purchasing but for the timing of the next renewal, and this should also be discussed. The problem is the balance between risk and cost. To obtain a budget, reasonable rationales should be presented. Here we predicted the estimated annual repair cost based on the history of repairs. Predictions of the repair cost based on field data and its use are reported.

  • Session ID: 2019-JP-25MP-D4

    Room D

    Tablets are an important dosage form accounting for about 50% of the administration form of drugs. Tablets are manufactured by consolidating powder or granules between a mortar and a pestle, except in special cases. Therefore, stresses with various vectors remain in the tablets after removing the pressure. The residual stress is known to be closely involved in tableting problems, such as capping and lamination. It may also have great impacts on significant characteristics related to tablet quality. We modeled the consolidation process of powder or granules with the Drucker-Prager Cap model and estimated the residual stress in the tablets by the finite element method. Applying the generalized regression model (JMP Pro 14) to these stress values, tablet quality characteristics were precisely predicted allowing identification of stress sites that affect quality characteristics. This presentation introduces case studies on tablets of different formulas or shapes and discusses the possibilities and limits of the prediction of quality characteristics by the generalized regression model.

  • Session ID: 2019-JP-25MP-A2

    Room A

    When analyzing actual production data to investigate the failure cause, the partition analysis shows outstanding acute ability. Since the partition report categorizes the range of the generation of non-defective and defective products and the report corresponding to a leaf clarifies the rate of generation (generation rate) of defective products (non-defective products) in the variable range, the condition of defective product generation can be clearly identified. Since this analysis result is understandable even without statistical knowledge, it is easy to explain the analysis result to the customer or stakeholder. It is also an advantage that the interaction is clarified without consciously incorporating the interaction term into a model like regression analysis. Of course, the partition analysis is not almighty. Sometimes the interaction cannot be clarified. The comparison of the difference in profiles by multiple analysis techniques and the study of the validity of the model from the technical viewpoints are important, although this is not limited to failure cause analysis. In the situation where market-returning goods prediction may become an important problem directly connected to management, the power of the reliability prediction of JMP is amazing. Since the AICc criteria are shown and the optimum model is clarified, the prediction excluding wishful bias can be performed. In particular, Defective Subpopulation (DS) distributions model is very useful because it is often suitable for the actual situation.

  • Session ID: 2019-JP-25MP-D3

    Room D

    Reversed-phase HPLC is an important analytical methodology for the estimation of the efficacy and safety of low-molecular drugs. In general, an active ingredient is a mixture containing a certain degree of impurities and reversed-phase HPLC can be used to determine the amount of individual impurities. Thus, development and optimization of HPLC conditions to separate the various components with high reproducibility is a extremely important process in the quality research. Major factors affecting the separation include the pH of the aqueous buffer and organic solvents used as eluent, temperatures of the analytical columns, and flow rates of the eluent. If the analytes are ionic compounds, the pH and character of organic solvents have particularly great impacts on the achievement of separation. However, changes in the times that the components take in passing through the analytical column (retention time) are not linear to the factors of the pH or the character of organic solvents, and it is very difficult to predict the achievement of separation in a modeling manner. Thus, the speaker attempted an approach where the achievement of separation is inferred within a arbitrary range of eluent conditions based on the predicted retention times derived from neural network modeling.

  • Session ID: 2019-JP-25MP-A4

    Room A

    The L12 orthogonal table of Plackett-Burman is an experimental design where the number of experiment times is as small as 12, although it takes as many as 11 factors. Therefore, it is often used in screening tests to evaluate the main effect when it cannot analyze the interaction in the statistic model analysis. However, it conducts various combinations of tests and is a highly efficient program from the viewpoint of the combination evaluation because the covering ratio of the combination of three factors is 100% and that of four factors is approximately 75%. The method of creating an experimental design with a high success rate was proposed using this combination evaluation, and many cases were found. This presentation introduces a method for achieving results with the smallest number of experiments by incorporating the technical ability into the experimental design using three representative cases. We strongly recommend to people who worry that the number of experiments is large in a three-level experiment to refer to this method.

  • Session ID: 2019-JP-50MP-D1

    Room D

    Trap of Reliability Data Analysis and Multivariate Analysis That Is Desired to Be Used by Reliability

    Koichi Endo, Engineer, Discrete Semiconductor Quality & Reliability Engineering Dep., Toshiba Electronic Devices & Storage Corporation

    Motohisa Hirono, Senior Manager, Quality Assurance & Medical Affairs Department, Medical Imaging Business Center, Ricoh Company, Ltd.

    In the reliability in our country, the mainstream is the analysis of Weibull probability paper, and the Arrhenius model also utilizes the probability paper. In addition, it is rare that the multivariate analysis is used in the test using multiple acceleration factors. This report first introduces the misuse of the bathtub curve and its measures. Secondly it introduces how the group of reliability data analysis techniques in JMP is useful and powerful.

  • Session ID: 2019-JP-25MP-A3

    Room A

    Use of JMP® for Study of Optimization of Wakame Seaweed Seed and Seedling Production

    Yoichi Sato, Manager, Bio-resources Business Development Division, Riken Food Co., Ltd.

    Kouichi Kanematsu, General Manager, Development Division, SiM24 Co., Ltd.

    Issues involving the wakame seaweed industry in recent years are 1) the reduction of production due to climatic variations, 2) the aging population of producers, and 3) the workload in the severe labor conditions of the cool season. In such an environment, Riken Food studied the optimization of raising seeds and seedlings that can be said to be wakame seaweed seedlings in cooperative research with Panasonic and SiM24. In order to investigate the influence of the environmental conditions of seawater temperature, light color and light intensity on the wakame seaweed seeds and seedlings, as well as young leaves, with the DOE platform in JMP. As a result, the optimum condition was successfully narrowed down after about half a year. Based on this research result, Riken Food Yuriage Factory in Natori, Miyagi Prefecture, promotes the wakame seaweed seedling production project. In addition, the optimization study of other algae is also performed currently using a similar optimization technique. In this presentation, we will explain the main points and the specific flow for performing the optimization study using the customized plan of JMP by adding a demonstration to make it understandable for beginners.

  • Session ID: 2019-JP-25MP-D2

    Room D

    Usefulness of Definitive Screening Design in the Design of Mini-Tablet Products

    Shui Usuda, Graduated Student, University of Toyama

    The concept of Quality by Design (QbD) recently focused on the development of pharmaceutical products. QbD aims at identifying causal relationships between the design factors that constitute the product (e.g., ingredients and manufacturing process condition) and the characteristics to design the optimal manufacturing method for pharmaceutical products based on scientific evidence. In the practice of QbD, the use of the design of experiments is strongly recommended. The present study investigated the preparation of mini-tablet products that are recently focused on as pediatric products using the definitive screening designs (DSD) included in JMP. DSDs are a new method for the design of experiments (DOE) for screening because they are capable of evaluating the effects of interactions and square terms with fewer experiments than in the conventional DOE. In the experiments, model products were prepared under different conditions based on the DSD to evaluate the causal relationships between design factors and characteristics. The experiment results successfully identified the design factors that affected product characteristics, obtaining several findings useful in designing mini-tablet products. Accordingly, DSDs are also very useful in designing the formula for mini-tablet products.

  • Session ID: 2019-JP-25MP-B2

    Room B

    An internet research study of a paired comparison was performed using the JMP Model Comparison platform to study the characteristics of pollinosis drugs required by general consumers with history of treatment for pollinosis under a trade-off environment. Furthermore, in the data obtained here, a correspondence analysis was performed to identify trends in daily activities of general consumers and characteristics of pollinosis drugs they require. The JMP Model Comparison platform allows an analysis of the results from option experiments, market research, and is capable of identifying the attributes of products or services that consumers like. The present study could identify the characteristics of pollinosis drugs required by general consumers with history of treatment for pollinosis.

  • Session ID: 2019-JP-PO-P8

    As an approach by the Ministry of Education, Culture, Sports, Science and Technology, the national knowledge and learning state investigations are performed for the sixth grades of elementary schools and the third grades of junior high schools in April of every year. The purpose is said to maintain and improve the equal opportunity and the level of compulsory education. However, the number of reports is small with regard to the consideration of effective education based on those investigations and of transverse analyses using the data obtained from the other investigations. This study analyzed the influence of various factors on academic ability based on the correct answer rate for four subjects in the third grade of junior high schools by a prefecture obtained in 2017 using a multiple regression analysis. As a result, the homework implementation rate of junior high school students, the self-affirming rate of junior high school students, the rate of participation in cultural affair clubs, the family conversation rate of junior high school students, the attendance rate of elementary school pupils, the early rising score (minute), and the smartphone and PC loving score (%) were extracted as positive factors, and the breakfast taking rate of elementary school pupils and the divorce rate per population of 1,000 were extracted as negative factors.

  • Session ID: 2019-JP-PO-P7

    Consideration by Searching Concerning Trends of Time and Content in Criminal Information

    Saho Watanuki, First-Year Master's Student, Toyo University

    At the present time when many types of information are coming and going, the easy acquisition of information is realized as technology progresses. Because of this, various opinions and ways of thinking were brought into focus. It is considered that the tendency of crime with conspicuous features appears in many pieces of significant information. For example, in cruel crime shaking up the world and crimes caused by entertainers, such as a celebrity, many pieces of information are shared, and matters aside from the crime get into the news. The study showing significant changes of the point and directivity of information in psychology includes the preceding study concerning groundless rumors and demagogy. However, how the deviation of such information spreads for a crime and changes with the lapse of time is not clarified at all. Therefore, the purpose of this study is to consider the reaction after the lapse of time when a crime influencing the world occurred by using text mining.

  • Session ID: 2019-JP-PO-P2

    General Plan for Screening and Modeling Experiments in Practical Business

    Akira Ogawa, Visiting Scholar, Mejiro University

    In the experimental design method, it is important to narrow down factors with a small number of experiment times, then model the characteristics using the narrowed down factors and determine the level in which the preferable characteristic is obtained. As a statistical quality control technique, the former and the latter are called screening and modeling, respectively. They are used by distinguishing each process. However, both are often implemented as a series of approaches in practical business, and the screening and modeling experiments should be planned from the general viewpoints in order to achieve the purpose of the experiment. This poster reviews the experimental design of screening and modeling in JMP as applied process improvement and its analysis results. As the point of the review, it is necessary to bring the specific target, the constraint condition to be considered and the recommended plan into perspective concerning the drafting of the general experimental design. Furthermore, after conducting the experiments, one should assess the target attainment and the cost effectiveness, as well as stakeholder evaluation. This presentation summarizes the issues from the general viewpoints of the experimental design and shows the prospect for the experimental design to be hereafter implemented.

  • Session ID: 2019-JP-PO-P10

    Practice of Factfulness Using JMP®

    Naohiro Masukawa, JMP Technical Engineer, SAS Japan

    Factfulness is a practice to overcome assumptions and look at the world correctly based on data. A book titled Factfulness proposes this concept and became the bestseller. This book provides easy-to-understand explanations on why we have wrong assumptions about the global issues of population, education and poverty based on data. The authors of the book have established the Gapminder foundation to publish data on statistics around the world for free. Thus, I practiced factfulness in my own way using the various functions for graph generation and analyses in JMP using the data. The results of the practice are published in JMP Public (public.jmp.com) effectively using the Dynamic Bubble Plot developed by the Gapminder Project, in addition to this presentation.

  • Session ID: 2019-JP-PO-P4

    Prediction of Rate of Failure by Quintile Regression

    Seiichi Yasui, Junior Associate Professor, Tokyo University of Science

    It is generally accepted that the prediction of the rate of failure from quality characteristic data is performed by the logistic regression. Since the logistic regression assumes logistical distribution for the rear distribution, a probit model assuming the normal distribution can be considered from this viewpoint. However, since such method requires data distribution and it is difficult to consider that the distribution of quality characteristics is always stable in the scene of quality improvement such as the reduction of the rate of failure, the robustness becomes a problem. Now we introduce the prediction of the rate of failure by quintile regression as a parametric prediction method. The quintile regression does not especially require the assumption of a distribution because it is a regression model against the quintile of the conditional distribution of quality characteristic data in a certain production condition. In addition, since the rate of failure is the rate of the quality characteristic data less than a certain quintile of the distribution or larger than the quintile, the quintile regression is the prediction of the rate of failure itself. JMP Pro can easily implement the quintile regression as a generalized regression. This presentation introduces the prediction of the rate of failure by the quintile regression through the example of the consideration of the relationship between the rate of failure and the production condition using the quintile regression in the production and the quality control of flame-retardant wood.

  • Session ID: 2019-JP-PO-P9

    Heat-and-eat curry is known as highly marketable processed food, with numerous products circulated in the market intensifying sales competition. For its sales strategy, the package design is a very important element, so the purpose of this study was to study the development of package for heat-and-eat curry that sells well. Analysis object is decided to be the questionnaire data concerning heat-and-eat curry package, then the principal component analysis and the selection regression analysis were performed, and the conjoint analysis of extracted attribute and level was performed using the L8 orthogonal table. As a result, we prepared an ideal package design draft by incorporating seven items: high degree of shop preference, mild hotness, high degree of recognition, display by photograph, innovative impression, light color package and display of quantity obtained as the elements of heat-and-eat curry package that sells well. The questionnaire created from the analysis results shows high support that 85% of respondents selected want to purchase and 91% selected want to eat, and the average price which respondents answered to the question “How much can you pay for this heat-and-eat curry?” resulted in \436.2.

  • Session ID: 2019-JP-PO-P6

    Proposal of New Synthesis Condition Searching Technique Using JMP®

    Yu Kitamura, Student, Kwansei Gakuin University

    Since the synthesis of the metal-organic framework (MOF) consists of multiple components, the slight difference of reaction condition results in passing through various states of reaction, and various crystal polymorphisms with different crystal structure is generated. However, since the parameters for synthesis are countless, and these elements affect the synthesis integrally, the evaluation is difficult. In order to synthesize the polymorphous crystal selectively, the experiments based on trial and error were essential. Therefore, we considered the use of the decision tree learning (partition), a kind of machine learning significantly developing in recent years, to allow a statistical evaluation of parameters. When we collect experiment data under the various synthesis conditions to create the partition, we successfully extracted the parameter (dominant factor) that mostly affects the synthesis. As a result, we successfully determined the synthesis condition of a new compound and demonstrated the usefulness of the research technique utilizing JMP.

  • Session ID: 2019-JP-PO-P3

    In recent years, online investigations using the internet are frequently performed for the various investigations implemented by enterprises. For example, even when implementing the national basis questionnaire investigation for the number of people exceeding 10,000, the use of the online investigation is effective from the viewpoints of economic efficiency and accessibility. In order to consider the budget for the investigation to be implemented and the burden of respondents, narrow down the number of question items to the minimum, and create the questionnaire with the essential contents corresponding to the purpose of the investigation, it is preferable to closely examine the question items and perform preliminary investigation based on a conceptual diagram and a characteristic diagram in advance. In addition, in case of large-scale online questionnaire investigation, a statistical approach by careful stratification and multivariate analysis is required in the analysis of obtained data without performing descriptive statistics only. This enables the detailed identification of the actual state of investigation objects and the introduction of the direction of reasonable proposals. This presentation uses a practical example of the large-scale online questionnaire investigation. Using this example, the idea of an investigation, the questionnaire design after preparation, the effective data acquisition by preliminary investigation, and the main one, the utilization of partition analysis and selection type multi-group principal component regression using JMP, and their results will be reported.

  • Session ID: 2019-JP-PO-P5

    Dissociation ranges from non-pathological experiences of imagination, devotion and immersion in normal individuals to pathological dissociation experiences of depersonalization, amnemonic experiences and dissociative identity disorder. It is necessary to identify the pathological components of dissociation for early intervention in serious psychotic disorders. Particularly in consultation rooms of the educational fields of junior high school, high school and colleges, many have a history of bullying and abuse, and the selection of the medical institution for referral should be considered based on is the severity of the dissociation in question. Therefore, the author prepared the Simplified Dissociation Scale of three-factor version consisting of 37 items that allows screening of dissociation in non-medical organizations. However, there was a problem that both exploratory and confirmative factor analyses were performed with the same data from college students while the scale was being prepared. To improve the accuracy of the scale, the structure of factors should be reviewed in different populations. In this presentation, therefore, a research study was performed using this scale and the psychotic health indices mainly in adolescent subjects of college/junior college students, and I would like to review the expressions of items, factor structure and appropriateness of the scale.

  • Session ID: 2019-JP-PO-P1

    Thermal Transportation Modeling of Nuclear Fusion Plasma Utilizing JMP® (Part 2)

    Masayuki Yokoyama, Professor, National Institute for Fusion Science

    In last year’s presentation, we showed the thermal transportation modeling of nuclear fusion plasma utilizing JMP. Thereafter, we developed various measures, such as the promotion of the utilization and mastery of JMP and the expansion of the explanation variables to 1) realize improvements in the coefficient of determination for a multiple regression analysis (improvement of data reproducibility ⇒ engineering utilization for operation control) and 2) to attempt the physical interpretation of regression results through variable selection (utilization of Akaike and Bayes information criteria) as the contact point from the aspect of a physical study. This year, we would like to report the development of these measures, the utilization of JMP in the physical field, and the current status of the improvement in the degree of recognition of the statistical analysis technique.

Back to Top