A aplicação de redes neuronais na deteção da influência do High Frequency Trading na negociação de ações (caso português)

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão da Informação

Mapping Mediterranean Portugal: Pastoral and Counter-pastoral

This article focuses on the different images of Mediterranean Portugal developed by three important Portuguese social scientists of the 20th century: the geographer Orlando Ribeiro, the ethnologist Jorge Dias and the social anthropologist José Cutileiro. The article argues that these different images stem from different ideological attitudes towards the countryside, ranging from pastoral to counter-pastoral, and are also related to different ways of addressing the links between the countryside and national identity.

Enhancing and implementing a pairs trading strategy

This paper designs a pairs trading model with the intent to identify existing profitable market opportunities to invest, i.e. traditionally strong correlated stocks that have diverged from its historical norm. It comprises a broad literature review on this strategy whose relevant findings (strategy improvements) are contemplated in the model. The authors combine the statistical results of the model with a backtesting analysis in order to provide guidance on the best investment opportunities.

The patent expiry as a major inflection point in a branded Rx drug's life and strategies to counter generic competition - exemplified at Pfizer's blockbuster drug Lipitor in the USA

A large number of expensive, but highly profitable branded prescription drugs will go off-patent in the USA between 2011 and 2015. Their revenues are crucial to fund the immense costs associated with the development of an innovative drug. The rising cost pressure on pharmaceutical stakeholders has increased the demand for more affordable medications, as provided by the branded drug's generic counterpart. Yet, research based incumbents are moving beyond the traditional late lifecycle strategies and deploy more aggressive tactics in order to protect their brands, as seen with Pfizer's Lipitor!. It is doubtful, whether these efforts will help the blockbuster business model to resist current market conditions.

Market risk charge of the trading book: a comparison of the Basel II and Basel III

This paper aims to investigate if the market capital charge of the trading book increased in Basel III compared to Basel II. I showed that the capital charge rises by 232% and 182% under the standardized and internal model, respectively. The varying liquidity horizons, the calibration to a stress period, the introduction of credit spread risk, the restrictions on correlations across risk categories and the incremental default charge boost Basel III requirements. Nevertheless, the impact of Expected shortfall at 97.5% is low and long term shocks decrease the charge. The standardized approach presents advantages and disadvantages relative to internal models.

Credit default swap (CDS) prediction model & trading strategy

This project focuses on the study of different explanatory models for the behavior of CDS security, such as Fixed-Effect Model, GLS Random-Effect Model, Pooled OLS and Quantile Regression Model. After determining the best fitness model, trading strategies with long and short positions in CDS have been developed. Due to some specifications of CDS, I conclude that the quantile regression is the most efficient model to estimate the data. The P&L and Sharpe Ratio of the strategy are analyzed using a backtesting analogy, where I conclude that, mainly for non-financial companies, the model allows traders to take advantage of and profit from arbitrages.

Pair trading: Clustering based on principal component analysis

This study focuses on the implementation of several pair trading strategies across three emerging markets, with the objective of comparing the results obtained from the different strategies and assessing if pair trading benefits from a more volatile environment. The results show that, indeed, there are higher potential profits arising from emerging markets. However, the higher excess return will be partially offset by higher transaction costs, which will be a determinant factor to the profitability of pair trading strategies. Also, a new clustering approach based on the Principal Component Analysis was tested as an alternative to the more standard clustering by Industry Groups. The new clustering approach delivers promising results, consistently reducing volatility to a greater extent than the Industry Group approach, with no significant harm to the excess returns.

Trading the differences in expectations between the CDS and bond markets

This paper uses the framework developed by Vrugt (2010) to extract the recovery rate and term-structure of risk-neutral default probabilities implied in the cross-section of Portuguese sovereign bonds outstanding between March and August 2011. During this period the expectations on the recovery rate remain firmly anchored around 50 percent while the instantaneous default probability increases steadily from 6 to above 30 percent. These parameters are then used to calculate the fair-value of a 5-year and 10- year CDS contract. A credit-risk-neutral strategy is developed from the difference between the market price of a CDS of the same tenors and the fair-value calculated, yielding a sharpe ratio of 3.2

O abuso de informação privilegiada (insider trading)

In the stock market, information takes on special relevance, due to the market’s permanent updating and the great fluidity of information existent therein. Just as in any other negotiations, the party with the better information has a bargaining advantage, as it is able to make more advantageous business decisions. However, unlike most other markets, the proper functioning of the stock market is greatly dependent on investors’ trust in the market itself. As such, if there are investors who, due to any condition they possess or office they hold, have access to relevant information which is not accessible to the general public, distrust is bred within the market and, consequently, investment is lessened. Thus, there is a need to prevent those who hold privileged information from using it in abusive ways. In Portugal, abuse of privileged information is set out and punished criminally in Article 378. of the Portuguese Securities Code (‘Código dos Valores Mobiliários’). In this dissertation, I have set out, firstly, to analyze the inherent conditions for there to be a crime of abuse of privileged information; secondly, to analyze two well-known cases, which took place and were decided in other jurisdictions, and attempt to understand how these cases would fall under Article 378. of the Portuguese Securities Code. Whereas the first case, Chiarella v. United States, was scrutinize under Article 378 of the Portuguese Securities Code, in the second, Lafonta v. AMF, the conclusion arrived at was that the crime taken place was different. This analysis allowed, on one hand, the application to a particular case of prerequisites and concepts which were explained, at a first approach, from a more theoretical perspective; on the other hand, it also allowed the further development of specific aspects of the regime, namely the difference between an insider and a tipee, as well as to more clearly set out the limits to the precise character of the information at hand.

Trading the changes of stock market index composition: Evidence from European markets

This paper studies the changes in European stock market indexes composition from 1995 to 2015. It was found that there are mixed price effects producing abnormal returns around the effective replacement of added and deleted stocks. The price pressure hypothesis seems to hold for added stocks in some indexes but not for deleted stocks as there is not a clear inversion of behaviour after the replacement. Finally, the building and back testing of a trading strategy aiming to capture some of those abnormal returns shows it yields a Sharpe Ratio of 1.4 and generates an annualised alpha of 11%.

Using artificial neural networks to generate trading signals for crude oil, copper and gold futures

In this thesis, a feed-forward, back-propagating Artificial Neural Network using the gradient descent algorithm is developed to forecast the directional movement of daily returns for WTI, gold and copper futures. Out-of-sample back-test results vary, with some predictive abilities for copper futures but none for either WTI or gold. The best statistically significant hit rate achieved was 57% for copper with an absolute return Sharpe Ratio of 1.25 and a benchmarked Information Ratio of 2.11.

A negociação no mercado a prazo do MIBEL

This master thesis has been developed during the internship in the Supervision Department of Supervision of the Intermediation and Market Structures of CMVM. My collaboration in such department was mainly focused on the derivatives market of the Iberian Electricity Market (MIBEL). MIBEL embodies two organized markets – the derivatives market in Portugal and the spot market in Spain The trading activity in the derivatives market of MIBEL is processed through the trading platform of the regulated market managed by OMIP, however, much of the negotiation is over-the-counter. The aim of this work is to describe the market from a legal and economic perspective and to analyse the evolution of the negotiation, namely the impact of OTC in the regulated market trading. To achieve this, I propose to analyse also MiFID and EMIR rules over derivative contracts and the role of central counterparties, as they both are important to the discussion. In parallel, we found that OTC transactions are considerably higher than those traded in the regulated market managed by OMIP, those findings can be justified by the contractual relationships based on trust already established between the partiesarties. Nevertheless, since 2011 this trend changed by an increase of the registered OTC. Thereafter, although the parties continued to trade bilaterally, these transactions were registered in a central counterparty in order to eliminate the inherent risks related to the OTC derivatives transactions. This change in the negotiation pattern may also be influenced by the mandatory reporting of transactions imposed by EMIR, that requires for some classes of derivatives the centralized clearing and for all other requires the implementation of risk mitigation techniques.

Actividade ígnea pós-paleozóica no continente português (elementos para uma síntese crítica)

The Author presents a synopsis about the post-Paleozoic igneous activity in continental Portugal. Subvolcanic massifs of Sintra, Sines and Monchique and the basaltic complex of Lisbon-Mafra are interpreted. The large network of dikes and sills occuring at north of Tagus river in Lisbon- Torres Vedras region as the dikes of Algarve and also those of diapiric formation are studied and compared. Also the doleritic dikes cuting the Hesperic Massif and the Great dike of Alentejo are studied. The Author presents an attempt of petrological and geochemical correlation-among these post-Paleozoic igneous rocks. For this more than 350 chemical analysis are used in order to elaborate several diagrams and some general conclusions are derived from them. The correlation between the origin of these igneous rocks and the opening of North Atlantic and the counter-clockwise rotation of the Iberia are also tried.

Automatic learning for the classification of chemical reactions and in statistical thermodynamics

This Thesis describes the application of automatic learning methods for a) the classification of organic and metabolic reactions, and b) the mapping of Potential Energy Surfaces(PES). The classification of reactions was approached with two distinct methodologies: a representation of chemical reactions based on NMR data, and a representation of chemical reactions from the reaction equation based on the physico-chemical and topological features of chemical bonds. NMR-based classification of photochemical and enzymatic reactions. Photochemical and metabolic reactions were classified by Kohonen Self-Organizing Maps (Kohonen SOMs) and Random Forests (RFs) taking as input the difference between the 1H NMR spectra of the products and the reactants. The development of such a representation can be applied in automatic analysis of changes in the 1H NMR spectrum of a mixture and their interpretation in terms of the chemical reactions taking place. Examples of possible applications are the monitoring of reaction processes, evaluation of the stability of chemicals, or even the interpretation of metabonomic data. A Kohonen SOM trained with a data set of metabolic reactions catalysed by transferases was able to correctly classify 75% of an independent test set in terms of the EC number subclass. Random Forests improved the correct predictions to 79%. With photochemical reactions classified into 7 groups, an independent test set was classified with 86-93% accuracy. The data set of photochemical reactions was also used to simulate mixtures with two reactions occurring simultaneously. Kohonen SOMs and Feed-Forward Neural Networks (FFNNs) were trained to classify the reactions occurring in a mixture based on the 1H NMR spectra of the products and reactants. Kohonen SOMs allowed the correct assignment of 53-63% of the mixtures (in a test set). Counter-Propagation Neural Networks (CPNNs) gave origin to similar results. The use of supervised learning techniques allowed an improvement in the results. They were improved to 77% of correct assignments when an ensemble of ten FFNNs were used and to 80% when Random Forests were used. This study was performed with NMR data simulated from the molecular structure by the SPINUS program. In the design of one test set, simulated data was combined with experimental data. The results support the proposal of linking databases of chemical reactions to experimental or simulated NMR data for automatic classification of reactions and mixtures of reactions. Genome-scale classification of enzymatic reactions from their reaction equation. The MOLMAP descriptor relies on a Kohonen SOM that defines types of bonds on the basis of their physico-chemical and topological properties. The MOLMAP descriptor of a molecule represents the types of bonds available in that molecule. The MOLMAP descriptor of a reaction is defined as the difference between the MOLMAPs of the products and the reactants, and numerically encodes the pattern of bonds that are broken, changed, and made during a chemical reaction. The automatic perception of chemical similarities between metabolic reactions is required for a variety of applications ranging from the computer validation of classification systems, genome-scale reconstruction (or comparison) of metabolic pathways, to the classification of enzymatic mechanisms. Catalytic functions of proteins are generally described by the EC numbers that are simultaneously employed as identifiers of reactions, enzymes, and enzyme genes, thus linking metabolic and genomic information. Different methods should be available to automatically compare metabolic reactions and for the automatic assignment of EC numbers to reactions still not officially classified. In this study, the genome-scale data set of enzymatic reactions available in the KEGG database was encoded by the MOLMAP descriptors, and was submitted to Kohonen SOMs to compare the resulting map with the official EC number classification, to explore the possibility of predicting EC numbers from the reaction equation, and to assess the internal consistency of the EC classification at the class level. A general agreement with the EC classification was observed, i.e. a relationship between the similarity of MOLMAPs and the similarity of EC numbers. At the same time, MOLMAPs were able to discriminate between EC sub-subclasses. EC numbers could be assigned at the class, subclass, and sub-subclass levels with accuracies up to 92%, 80%, and 70% for independent test sets. The correspondence between chemical similarity of metabolic reactions and their MOLMAP descriptors was applied to the identification of a number of reactions mapped into the same neuron but belonging to different EC classes, which demonstrated the ability of the MOLMAP/SOM approach to verify the internal consistency of classifications in databases of metabolic reactions. RFs were also used to assign the four levels of the EC hierarchy from the reaction equation. EC numbers were correctly assigned in 95%, 90%, 85% and 86% of the cases (for independent test sets) at the class, subclass, sub-subclass and full EC number level,respectively. Experiments for the classification of reactions from the main reactants and products were performed with RFs - EC numbers were assigned at the class, subclass and sub-subclass level with accuracies of 78%, 74% and 63%, respectively. In the course of the experiments with metabolic reactions we suggested that the MOLMAP / SOM concept could be extended to the representation of other levels of metabolic information such as metabolic pathways. Following the MOLMAP idea, the pattern of neurons activated by the reactions of a metabolic pathway is a representation of the reactions involved in that pathway - a descriptor of the metabolic pathway. This reasoning enabled the comparison of different pathways, the automatic classification of pathways, and a classification of organisms based on their biochemical machinery. The three levels of classification (from bonds to metabolic pathways) allowed to map and perceive chemical similarities between metabolic pathways even for pathways of different types of metabolism and pathways that do not share similarities in terms of EC numbers. Mapping of PES by neural networks (NNs). In a first series of experiments, ensembles of Feed-Forward NNs (EnsFFNNs) and Associative Neural Networks (ASNNs) were trained to reproduce PES represented by the Lennard-Jones (LJ) analytical potential function. The accuracy of the method was assessed by comparing the results of molecular dynamics simulations (thermal, structural, and dynamic properties) obtained from the NNs-PES and from the LJ function. The results indicated that for LJ-type potentials, NNs can be trained to generate accurate PES to be used in molecular simulations. EnsFFNNs and ASNNs gave better results than single FFNNs. A remarkable ability of the NNs models to interpolate between distant curves and accurately reproduce potentials to be used in molecular simulations is shown. The purpose of the first study was to systematically analyse the accuracy of different NNs. Our main motivation, however, is reflected in the next study: the mapping of multidimensional PES by NNs to simulate, by Molecular Dynamics or Monte Carlo, the adsorption and self-assembly of solvated organic molecules on noble-metal electrodes. Indeed, for such complex and heterogeneous systems the development of suitable analytical functions that fit quantum mechanical interaction energies is a non-trivial or even impossible task. The data consisted of energy values, from Density Functional Theory (DFT) calculations, at different distances, for several molecular orientations and three electrode adsorption sites. The results indicate that NNs require a data set large enough to cover well the diversity of possible interaction sites, distances, and orientations. NNs trained with such data sets can perform equally well or even better than analytical functions. Therefore, they can be used in molecular simulations, particularly for the ethanol/Au (111) interface which is the case studied in the present Thesis. Once properly trained, the networks are able to produce, as output, any required number of energy points for accurate interpolations.

Matching CO2 large point sources and potential geological storage sites in mainland Portugal

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para obtenção do grau de Mestre em Engenharia do Ambiente, Perfil Gestão e Sistemas Ambientais