UID:
almahu_9949984245102882
Format:
1 online resource (464 pages)
Edition:
Second edition.
ISBN:
9780443138300
Content:
This book, authored by Erwin B. Montgomery Jr., explores the intricacies of biomedical research and the statistical challenges associated with it. It provides a comprehensive analysis of traditional and contemporary research methodologies, emphasizing the role of machine learning and artificial intelligence in advancing scientific inquiry. The book discusses the importance of precision, accuracy, and causality in experimental designs, while addressing the limitations and potential fallacies in scientific reasoning. Through various case studies, it highlights the practical applications and theoretical underpinnings of biomedical, clinical, and computational research. Intended for researchers and practitioners in the field, it aims to enhance understanding and improve practices in biomedical research.
Note:
Front Cover -- Reproducibility in Biomedical Research -- Copyright Page -- Dedication -- Quotes -- Contents -- Preface to the second edition -- Looking just over the horizon -- Machine learning artificial intelligence -- Translational research -- Biological realism and Chaos and Complexity -- Probability and statistical epistemology -- Randomness as fundamental and foundational -- Finally… -- Preface to the first edition -- 1 Introduction -- Productive irreproducibility -- The multifaceted notion of reproducibility and irreproducibility -- Beyond "stamp collecting" -- Replication reproducibility -- Within-experiment reproducibility -- Narrow, broad, conceptual, and translational reproducibility -- Turning from the past with an eye to the future -- Machine learning artificial intelligence -- A challenge to clinical research posed by a new notion of translational research -- Toward more realistic experimental designs and analyses -- The fundamental causes of unproductive irreproducibility -- Illogic -- The paradox and dilemma of the "population" -- The logic of population and sample -- The proof of the pudding, in this case sample/population biomedical research -- Wrestling with the fundamental problem -- Proceeding from what is certain but not useful to what is uncertain but useful -- Precision versus accuracy -- Precision and statistical distributions -- Combinatorics -- Dynamics -- Machine learning artificial intelligence and the emperor's new clothes -- Knowledge is prediction, prediction is reproducibility or productive irreproducibility -- The ascendancy of experimental science -- Challenges to prediction and thus biomedical research -- Predicting to the unknowable population -- Hume's Problem of Induction -- When traditional experimental design and statistics breed unproductive irreproducibility -- Data do not and cannot speak for themselves.
,
Reductionism and the fundamental problem -- Summary -- 2 The problem of irreproducibility -- Getting a handle on the scope of unproductive irreproducibility -- Type I versus type II errors -- Precision versus accuracy -- Sensitivity versus specificity -- Logic -- The inescapable risk of irreproducibility -- Institutional responses -- Who speaks for reproducibility and irreproducibility? -- Fundamental limits to reproducibility as traditionally defined -- Variability, central tendency, Chaos, and Complexity -- Summary -- 3 Validity of biomedical science, reproducibility, and irreproducibility -- Science must be doing something right and therein lies reproducibility and productive irreproducibility -- Legacy of injudicious use of scientific logical fallacies -- The importance of past accomplishments and the promise of future ones -- Science versus human knowledge of it -- The necessity of enabling assumptions -- The Central Tendency -- The Scientific Method -- Reductionism -- Causality -- Special cases of irreproducible reproducibility -- Science as inference to the best explanation -- Summary -- 4 The logic of certainty versus the logic of discovery -- Certainty, reproducibility, and logic -- Unpacking certainty -- Establishing certainty -- Logic as proof -- Deductive logic-certainty and limitations -- Propositional logic -- Syllogistic deduction -- Foundations of syllogistic deduction -- Centrality of syllogistic deduction and the Fallacy of Four Terms in biomedical research -- Epistemic risk -- Judicious use of the Fallacy of Four Terms -- Metaphors and syllogisms -- Partial, probability, practical, and causal syllogisms -- Practical syllogism -- Causal syllogism -- Probability and statistical syllogisms -- Induction -- Inductive inferences -- Method of difference -- Method of agreement -- Joint method of agreement and difference.
,
Method of residual -- Method of concomitant variation -- The Duhem-Quine thesis -- Summary -- 5 The logic of probability and statistics -- Probability has always been central, statistics only relatively recently -- Precision versus accuracy, epistemology versus ontology -- The purpose of the chapter -- Continuing legacy of notions of probability -- The value of the logical perspective in probability and statistics -- Metaphysics: ontology versus epistemology and biomedical reproducibility -- Central Tendency versus variance as the true measure of reality -- Probability -- Empirical probability -- Hume's Problem of Induction -- Probability calculus -- Bayes' theorem -- Positive and Negative Predictive Value -- Receiver Operator Characteristic Curve analyses -- Statistics -- Normal distribution -- Use of normal distribution and measurement -- Constructing an artificial standard normal distribution in experimental design -- Correlational analysis -- Mathematical optimization -- Key general assumptions whose violation risks unproductive irreproducibility -- Summary -- 6 Causation, process metaphor, and reductionism -- Renewed need for causation -- Practical syllogism and beyond -- Centrality of hypothesis to experimentation and centrality of causation to hypothesis generation -- Ontological sense of cause -- Principle of Causational Synonymy -- Principle of Informational Synonymy -- Reductionism and the Fallacies of Composition and Division -- Other fallacies as applied to cause -- Discipline in the Principles of Causational and Informational Synonymy -- Process metaphor -- Summary -- 7 Case studies in clinical biomedical research -- Forbearance of repetition -- Purpose of clinical research as the standard -- Clinical research as an epistemic necessity for translational research -- Clinical importance.
,
Sensitivity and specificity in choosing α cutoff values -- Establishing clinical importance -- Specific features to look for in case studies -- Case study-two conflicting studies of hormone use in postmenopausal women, which is irreproducible? -- Consequences -- Why the dominance of the Women's Health Initiative Study over the Nurses' Health Study? -- Aftermath -- Summary -- 8 Case studies in basic biomedical research -- Forbearance of repetition -- Purpose -- Setting the stage -- The value of a tool from its intended use -- What is basic biomedical research? -- Scientific importance versus statistical significance -- Reproducibility and the willingness to ignore irreproducibility -- Specific features to look for in case studies -- Case study-pathophysiology of parkinsonism and physiology of the basal ganglia -- Summary -- 9 Case studies in computational biomedical research -- Theorizing versus computational modeling with simulation -- Scope of computation in biomedical research -- Importance of mathematical and computational modeling and simulations -- The notion of irreproducibility in mathematical and computational modeling and simulations -- Sources of irreproducibility in the narrow sense -- Compilation versus runtime errors -- Complexity and Chaos and underdetermination in computational modeling and simulations -- The necessity of biological constraints and the fallacy of confirming the consequence -- Setting the stage -- Computational importance -- Specific features to look for in mathematical and computational studies -- Case studies -- Summary -- 10 Case studies in translational research -- Translational research as the ultimate goal of basic and clinical research -- Contemporary perspective on translational research -- Definitions of basic, clinical, and translational research -- Unproductive irreproducibility in translational research.
,
Translational research and clinical practice -- Differences among basic, clinical, and translational research -- Dealing with variance -- The Large Number theorem in translational research -- n of 1 Problem -- n of 1 Studies as research rather than routine medical care -- Formalization of n of 1 trials -- Basket trials -- Case series -- Information loss -- No "free lunch" -- Preserving information -- Multidimensional representation -- Sample-centric clinical research -- When discussing the tools, the carpenter cannot be excluded -- Uniqueness and challenge of translational research -- Effects of basic and clinical research key personnel -- Ethical value -- Different criteria for diagnoses -- Case study -- Evidence-Based Medicine and translational research -- Summary -- 11 Case studies in machine learning artificial intelligence -- The current environment of machine learning artificial intelligence -- Machine learning AI in the context of biomedical research -- Types of machine learning AI -- Learned versus learning machine learning AI -- Example of a neural network machine learning AI -- The game of warmer/colder -- Difference between machine learning AI and other analytic methods -- The notion of models, empirical and analytic probabilities -- Similarities between machine learning AI and other methods, such as regression analyses -- Least summed squared error and best fit -- The problem of mathematical optimization -- Quality assurance in multivariate regression and implications for machine learning AI -- Analysis of residuals -- The nature and size of the sample or training set -- Fallacy of Four Terms -- What is the purpose or goal? -- To whom or what is the machine learning AI algorithm to apply? -- How should one select the training set? -- What is the learning methodology? -- The notion of error -- Only as good as the "gold standard".
,
Error analyses as quality control.
Additional Edition:
ISBN 9780443138294
Language:
English
Bookmarklink
