Machine learning has revolutionized many fields, including science, healthcare, and business. It is also widely used in network data analysis. This chapter provides an overview of machine learning methods and how they can be applied to network data. Machine learning can be used to clean, process, and analyze network data, as well as make predictions about networks and network attributes. Methods that transform networks into meaningful representations are especially useful for specific network prediction tasks, such as classifying nodes and predicting links. The challenges of using machine learning with network data include recognizing data leakage and detecting dataset shift. As with all machine learning, effective use of machine learning on networks depends on practicing good data hygiene when evaluating a predictive model’s performance.

Chapter 3 provides an overview of all elements of the predictive modeling process, from the selection of training and test data sets, parallel multivariate feature selection experiments and deciding on an optimal multivariate biomarker, to building, tuning, validating, and testing predictive models implementing the optimal biomarker. Discussed are also such topics as bias-variance tradeoff, segmentation models, and committees of predictive models.

Chapter 10 covers the random forests algorithm for classification. Presented are also the impurity metrics applicable to splitting nodes in classification trees (Gini, entropy, and misclassification impurity), as well as permutation-based and impurity-based variable importance measures.

Chapter 17 considers Goethe’s extensive collections, which ranged in subject matter from art and ethnography to natural history and scientific instruments, and also included a vast library. It uses the period around his Italian journey (1786–8), when his involvement with art and art objects was particularly intense, to highlight tensions within his approach to collecting which apply throughout his career as a collector. The chapter also addresses the complexity of classifying Goethe’s collections, owing to their scale and diversity, and to the variety of his own collecting habits.

The concept of abnormal mood has been a matter of a millennia-long debate in philosophy and medicine, while the diagnosis and classification of mood disorders remains a complex and controversial issue even in modern psychiatry. A centrepiece of this debate is the conceptualisation of mood and, by extension, mood disorders as a multi-dimensional spectrum with transdiagnostic symptoms (i.e., a continuous diagnostic classification) or as discrete nosological entities (i.e., a categorical diagnostic classification). Theoretical models and arguments based on empirical evidence have been proposed for both the distinct categorisation of abnormal mood states and the affective continuum perspective, which may also encompass psychosis and psychotic disorders. Although the conceptualisation of mood as a spectrum ranging from unipolar depression to unipolar mania may be the most suitable, this approach requires further evidence before it can replace the categorical classifications firmly employed in clinical practice for more than a century.

In his book General Psychopathology, first published in 1913, Jaspers presented a methodological framework for exploring the phenomenology of symptoms of psychiatric disorders as well as relating experimental psychology and nosology to phenomenology. This chapter briefly introduces the phenomenological approach to symptoms and how this has influenced symptom- as opposed to diagnostic criterion-based assessment instruments, such as those based on the diagnostic statistical manual. A transcultural and historical perspective is employed to identify relevant symptoms of mood disorders and their temporal course. Descriptions and definitions of classical symptoms are provided and extended based on modern evidence to include changes in self-imagery, moral emotions, self-blame-related action tendencies, as well as mood-congruent biases in the representation of the past and future. Lastly the contribution of psychopathology to future subsyndrome discovery, translational cognitive neuroscience, and network-based approaches to the psychopathology of mood disorders is discussed.

‘Psychotic disorders’ is an umbrella term for psychiatric conditions featuring psychosis, including mood disorders. Despite the prominence of psychotic symptoms across the psychotic spectrum, a distinction between schizophrenia and affective psychoses has been historically established. Findings from genetic studies support the aetiological overlap between affective and non-affective psychosis, although poor characterisation of the schizoaffective population still poses a challenge. Likewise, literature points to shared environmental risk factors between bipolar disorder and schizophrenia. Neuroimaging evidence suggest significant similarities in the pathophysiology of the brain between affective and non-affective psychosis. An overlap is also observed in other biological and behavioural illness markers, as well as in the pharmacotherapy of psychotic disorders. Current diagnostic entities may not accurately delineate the aetiology and pathophysiology of these conditions. Modern classification approaches, such as the RDoC framework, propose the adoption of aetiological factors and pathophysiological evidence to characterise patients, rather than categorical diagnoses based on symptoms.

In many applications, dimensionality reduction is important. Uses of dimensionality reduction include visualization, removing noise, and decreasing compute and memory requirements, such as for image compression. This chapter focuses on low-rank approximation of a matrix. There are theoretical models for why big matrices should be approximately low rank. Low-rank approximations are also used to compress large neural network models to reduce computation and storage. The chapter begins with the classic approach to approximating a matrix by a low-rank matrix, using a nonconvex formulation that has a remarkably simple singular value decomposition solution. It then applies this approach to the source localization application via the multidimensional scaling method and to the photometric stereo application. It then turns to convex formulations of low-rank approximation based on proximal operators that involve singular value shrinkage. It discusses methods for choosing the rank of the approximation, and describes the optimal shrinkage method called OptShrink. It discusses related dimensionality reduction methods including (linear) autoencoders and principal component analysis. It applies the methods to learning low-dimensionality subspaces from training data for subspace-based classification problems. Finally, it extends the method to streaming applications with time-varying data. This chapter bridges the classical singular value decomposition tool with modern applications in signal processing and machine learning.

Personality disorder represents a diagnosis very different from others in psychiatry. This is because it describes a long-standing integral part of a person, not just an affliction that has happened. Because of the sensitivity of ascribing a core part of a person’s being to the impersonality of a diagnostic term, the subject has been widely stigmatised. However, the condition is very common and affects one-tenth of the population. In this chapter, the clinical features of personality disorder identified in the new ICD-11 severity classification are described and their value illustrated. A fuller description of the ICD-11 classification can be found in another College publication.

There are five levels of diagnosis of personality disorder, including the sub-syndromal form – personality difficulty – which is by far the most common. The diagnosis of borderline personality disorder is the most used in practice but is a heterogeneous term that overlaps with almost every other disorder in psychiatry. All personality disorders have approximately equal genetic and environmental precursors, and the involvement of childhood adverse experiences and trauma is unfortunately true for this as for all psychiatric disorders.

Depressive disorders have been recognised since antiquity, although how they have been described and understood has changed considerably over time. In this chapter, we outline key aspects of the history of depression as well as some of the limitations in its current classification in ICD-11 and DSM-5. We describe the range of symptoms experienced in depressive disorders, together with the recognised variations in clinical presentation and how these are conceptualised and classified. The relationship between depression and related disorders including anxiety disorders, premenstrual dysphoric disorder and grief is discussed, as well as boundary issues with bipolar disorder and primary psychotic disorders. We review current knowledge about depression’s considerable psychiatric and medical comorbidity, along with its epidemiology, natural history and health burden. A brief practical guide to assessing depressive disorders is given, together with rating scales that are useful for clinical assessment and monitoring.

Today’s platformization of services was an “unforeseen” phenomenon when the WTO was established. However, through the pro-liberalization WTO jurisprudence developed in past decades, the GATS market access commitments have played more than a marginal role in the story of the emergence of datafication, leaving the door open for big tech companies. Without question, platforms’ associated datafication practices influence human society, in terms of their beauty as well as their peril. The perilous side therefore calls into question whether international trade commitments negotiated in the pre-digital age should be sustained in this datafied world. In the FTA context, Chapter 3 argues that listing nonconforming measures under the Cross-Border Trade in Services Chapter represents the most critical step in reserving a state’s “digital sovereignty.” Such inscriptions ensure that a state does not risk committing to future services and associated delivery means that did not exist at the time of treaty negotiations. Nonetheless, market access obligations do not prevent states from adopting domestic regulations in pursuit of legitimate national policies. Digital trade liberalization must be accompanied by the introduction of data regulations to address the potential risks and harms.