Note: Intro -- Dedication -- Foreword -- Preface -- Acknowledgments -- Contents -- Contributors -- About the Editors -- Part I: Introduction -- Chapter 1: Machine Learning in Wildlife Biology: Algorithms, Data Issues and Availability, Workflows, Citizen Science, Code Sharing, Metadata and a Brief Historical Perspective -- 1.1 Introduction -- 1.2 Some Terminology -- 1.3 A Few Paragraphs on the History of Machine Learning -- 1.4 Machine Learning in Ecology and Wildife Biology to Date -- 1.5 Algorithms as a Bottleneck for Wildlife Conservation -- 1.6 Data Issues and Availability Related to Data Mining and Machine Learning -- 1.7 Workflows -- 1.8 Citizen Science -- 1.9 A Great Future could be around the Corner, Waiting for you Online, and in the Wilderness of this World -- References -- Chapter 2: Use of Machine Learning (ML) for Predicting and Analyzing Ecological and 'Presence Only' Data: An Overview of Applications and a Good Outlook -- 2.1 Introduction -- 2.2 Popular and Widely Available Machine Learning Techniques -- 2.3 Applications of Machine Learning in Wildlife Biology -- 2.4 Strengths and Some Described Weaknesses of Machine Learning -- 2.5 A Case Example -- 2.6 Machine Learning in Climate Change Models and Other Complex Applications -- 2.7 Conclusions: Future Outlook and Topics Awaiting Research and Application for Machine Learning (ML) -- References -- Chapter 3: Boosting, Bagging and Ensembles in the Real World: An Overview, some Explanations and a Practical Synthesis for Holistic Global Wildlife Conservation Applications Based on Machine Learning with Decision Trees -- 3.1 Introduction -- 3.2 A Quick Refresher on Linear Models (LMs), Parsimony and Classification and Regression Trees (CARTs) -- 3.3 Boosting -- 3.3.1 What Boosting is in a Nutshell -- 3.3.2 Short History of 'Boosting' -- 3.3.3 Why is Boosting so Powerful? -- 3.4 Bagging. , 3.4.1 What Bagging is in a Nutshell -- 3.4.2 Short History of Bagging -- 3.4.3 Why Bagging is so Powerful -- 3.5 Ensemble Models -- 3.5.1 What is an Ensemble Model? -- 3.5.2 History of Ensemble Models -- 3.5.3 Why Ensemble Models are so Powerful -- 3.6 Model Applications and Inference -- 3.6.1 Boosting Experiences and Applications -- 3.6.2 Bagging Experiences and Applications -- 3.6.3 Ensembles -- 3.6.4 Precautionary, Pro-Active, and Predictive Models for Better Resource Conservation Management -- 3.7 A Commonly Heard Criticism and Misunderstanding of Machine Learning, and Characteristics of Man-Made Science and Conservation Driven by Reductionism -- 3.8 Synthesis and Outlook -- References -- Part II: Predicting Patterns -- Chapter 4: From Data Mining with Machine Learning to Inference in Diverse and Highly Complex Data: Some Shared Experiences, Intellectual Reasoning and Analysis Steps for the Real World of Science Applications -- 4.1 Introduction -- 4.2 Model Selection with Many Predictors as an Analysis Scheme and as a Major Platform for Statistical Testing, Prediction and Inference -- 4.3 Confront Models with Data: Moving towards an Evidence-Based Analysis -- 4.4 A Real-World Data and Analysis Workflow Example -- 4.4.1 Pre-Assumption (as Experienced in Real World Examples) -- 4.5 Real World Tools and Minimum Approaches to Start Data Mining and Predictions -- 4.6 A Set of Commonly Heard Criticisms and Comments for these Data Mining Steps and How to Answer them from a Machine Learning Aspect for Best inferences -- 4.7 On 'Best Professional Practices', Professional Bias, Ignorance, Misconduct, Professional Societies, Education and Culture: What is a Lie and Punishable Intent when underlying Methods Problems are well-known but are ignored? -- 4.8 Conclusion -- References. , Chapter 5: Ensembles of Ensembles: Combining the Predictions from Multiple Machine Learning Methods -- 5.1 Introduction -- 5.2 Methods -- 5.2.1 Data Set -- 5.2.2 Accuracy Assessment -- 5.2.3 Modeling Algorithms -- 5.3 Results and Discussion -- References -- Chapter 6: Machine Learning for Macroscale Ecological Niche Modeling - a Multi-Model, Multi-Response Ensemble Technique for Tree Species Management Under Climate Change -- 6.1 Introduction -- 6.2 Controlling Bias and Variance -- 6.3 Ensemble Learning Via Decision Trees -- 6.4 Ensemble Models -- 6.4.1 Bagging, Random Forest and Extreme Random Forests -- 6.4.2 Boosting Decision Trees -- 6.4.3 RuleFit -- 6.5 Multiple Abundances - Habitat Suitability -- 6.6 Explanatory Variables (Predictors) -- 6.7 Multi-Model Ensemble Approach -- 6.8 Results and Interpretation -- 6.8.1 Model Reliability -- 6.8.2 Prediction Confidence -- 6.8.3 Combined Habitat Quality and Prediction Confidence -- 6.8.4 Predictor Importance -- 6.9 Discussion -- 6.10 Conclusion -- References -- Chapter 7: Mapping Aboveground Biomass of Trees Using Forest Inventory Data and Public Environmental Variables within the Alaskan Boreal Forest -- 7.1 Introduction -- 7.2 Methods and Materials -- 7.2.1 Biomass Data -- 7.2.2 Environmental Factors -- 7.2.3 The Calibration and Validation Datasets -- 7.2.4 Statistical Methods -- 7.2.5 Predictive Maps -- 7.3 Results -- 7.3.1 Variable Selection and Importance -- 7.3.2 Random Forests Analysis Model Assessment -- 7.3.3 Influence of the Environmental Factors on Aboveground Forest Biomass -- 7.3.4 Regression Tree Analysis Model Assessment -- 7.3.5 Spatial Dependency of Aboveground Forest Biomass -- 7.3.6 Predicted Aboveground Forest Biomass Patterns -- 7.4 Discussion -- References -- Part III: Data Exploration and Hypothesis Generation with Machine Learning. , Chapter 8: 'Batteries' in Machine Learning: A First Experimental Assessment of Inference for Siberian Crane Breeding Grounds in the Russian High Arctic Based on 'Shaving' 74 Predictors -- 8.1 Introduction -- 8.2 Methods -- 8.2.1 Presence and Absence Points for Siberian Crane -- 8.2.2 GIS Predictors -- 8.2.3 'Battery' Runs in Salford System's Predictive Modeler (SPM7) -- 8.3 Results -- 8.3.1 Predictive Performance Metrics -- 8.3.2 Visual Assessment of Prediction Maps -- 8.4 Discussion -- Appendix 1: Details of 74 GIS Environmental layers Used in the Model Prediction (+ 3 Additional Internal Columns) -- Appendix 2 -- List of Top 20 Predictors, as identified by TreeNet ranking -- Appendix 3 -- Prediction Model Details for the Best Performing Model (the 'Kitchen sink model' with 74 predictors) -- Appendix 4 -- Prediction Map 3 for the 'BIO14 model' -- Prediction Map 4 for the 'TMax12BIO14 model' -- Prediction Map 5 for the 'Top5 model' -- Prediction Map 6 for the 'Top10 model' -- Prediction Map 7 for the 'Top29 model' -- Prediction Map 8 for the 'Top35 model' -- Prediction Map 9 for the 'Bottom 44 model' -- Prediction map 10 for the 'Leaving out top 3 interacting predictors model' -- References -- Chapter 9: Landscape Applications of Machine Learning: Comparing Random Forests and Logistic Regression in Multi-Scale Optimized Predictive Modeling of American Marten Occurrence in Northern Idaho, USA -- 9.1 Introduction -- 9.2 Methods -- 9.2.1 Study Area -- 9.2.2 Occurrence Data and Logistic Regression Model -- 9.2.3 Predictor Variables for Analysis -- 9.2.4 Modeling Approaches -- 9.2.5 Model Assessment -- 9.3 Results -- 9.3.1 Random Forests Univariate Scaling -- 9.3.2 Random Forests Multivariate Model -- 9.3.3 Model Comparison -- 9.3.4 Model Performance -- 9.4 Discussion -- 9.5 Conclusion -- References. , Chapter 10: Using Interactions among Species, Landscapes, and Climate to Inform Ecological Niche Models: A Case Study of American Marten (Martes americana) Distribution in Alaska -- 10.1 Introduction -- 10.1.1 Stochastic Gradient Boosting -- 10.1.2 Variable Interactions -- 10.1.3 Ecological Niche Models -- 10.2 Methods -- 10.2.1 Training Dataset -- 10.2.2 Model Iterations -- 10.2.3 Interaction Network Graphs -- 10.2.4 Landscape Predictions -- 10.3 Results -- 10.3.1 Varclus Analysis -- 10.3.2 Full Model -- 10.3.3 Top Sub-Models -- 10.3.4 Other Sub-Models -- 10.3.5 Spatial Models -- 10.4 Discussion -- 10.4.1 Parsimonious Versus Highly Interactive Models -- 10.4.2 High Level Categorical (HLC) Predictors -- 10.4.3 Interaction Effects -- 10.4.4 Predicted Marten distribution in Alaska -- 10.4.5 Conclusions and Suggested Practices -- References -- Chapter 11: Advanced Data Mining (Cloning) of Predicted Climate-Scapes and Their Variances Assessed with Machine Learning: An Example from Southern Alaska Shows Topographical Biases and Strong Differences -- 11.1 Introduction -- 11.2 Methods -- 11.2.1 GIS Data and Operations -- 11.2.2 Data Mining -- 11.3 Results -- 11.4 Discussion -- Appendix A Example Map of Data Sets Used for this Machine Learning Assessment of Climate Models: Adaptwest in July. Raw Climate Surface and All GIS Maps are available from the Authors on Request -- Appendix B Remaining Details of the TreeNet Model not shown in the text: (a) gains curve, and partial dependence plots for (b) proximity to road, (c) proximity to river -- References -- Chapter 12: Using TreeNet, a Machine Learning Approach to Better Understand Factors that Influence Elevated Blood Lead Levels in Wintering Golden Eagles in the Western United States -- 12.1 Methods -- 12.2 Results -- 12.3 Discussion -- References. , Part IV: Novel Applications of Machine Learning Beyond Species Distribution Models.