Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2015-10-21T11:05:40Z No. of bitstreams: 1 PAULO A. SCHMIDT_.pdf: 1386765 bytes, checksum: f2caadfc119f6eda2a41f0afb9efe1f1 (MD5) / Made available in DSpace on 2015-10-21T11:05:40Z (GMT). No. of bitstreams: 1 PAULO A. SCHMIDT_.pdf: 1386765 bytes, checksum: f2caadfc119f6eda2a41f0afb9efe1f1 (MD5) Previous issue date: 2015-07-27 / Nenhuma / A previsão de preços de ações é um assunto de grande interesse tanto por parte de agentes de mercado quanto da comunidade científica e acadêmica. Ao mesmo tempo, o problema é considerado como um dos mais desafiadores no tratamento de séries temporais, dada sua natureza altamente dinâmica. Uma ampla gama de estudos propõe-se a abordar o tema. Alguns com resultados bastante promissores fazem uso de Redes Neurais Artificiais (RNAs) do tipo Focused Time Lagged FeedForward Network (FTLFN), as quais apresentam mecanismos de memória capazes de detectar padrões temporais. Em muitos casos, no entanto, as capacidades da rede neural não são devidamente exploradas, limitando-se a testes com um conjunto mínimo de parâmetros. Além disso, a maioria dos estudos de previsões de preços de ações possui como foco períodos de baixa frequência, como dias ou meses. Contudo, devido à facilidade de acesso à informação nas últimas décadas e à automatização das negociações em bolsas de valores, estas são realizadas cada vez mais sob horizontes de curto prazo, como horas, minutos ou segundos. Existe, portanto, a necessidade de se expandir o conhecimento em relação a previsões dentro deste cenário. Neste sentido, este trabalho tem como objetivo uma investigação das reais potencialidades de previsão das FTLFNs sobre preços de ações no período intradiário. Sua memória de curto prazo e tamanho de camada oculta são explorados de forma ampla e aprofundada, através dos quais se buscou identificar o impacto das diferentes configurações nos resultados de acurácia dentro do contexto considerado. Na tentativa de oferecer suporte a melhores previsões, analisa-se também a influência de indicadores da Análise Técnica sobre o modelo. De forma mais geral, procura-se ampliar o entendimento a respeito tanto das capacidades de previsão das redes do tipo FTLFN como de sua empregabilidade em séries temporais financeiras intradiárias, ainda pouco exploradas na literatura. Os resultados obtidos mostram que, assim como investidores humanos, também as FTLFNs são capazes de se beneficiar enormemente de padrões formados pelos históricos dos sinais de entrada, a fim de prover previsões de maior qualidade dentro do contexto proposto neste trabalho. O mesmo não pode ser afirmado a respeito dos indicadores da Análise Técnica escolhidos, uma vez que em sua grande maioria aumentam os erros de previsão. As evidências apresentadas baseiam-se em experimentações sobre diferentes conjuntos de sinais, oferecendo robustez às conclusões alcançadas e permitindo que a metodologia e os resultados sirvam como base para futuras pesquisas relacionadas a previsões dentro de cenários de alta frequência. / Stock price prediction is a subject of great interest for both market agents and scientific and academic community. At the same time, this problem is considered to be one of the most challenging in time series forecasting, due to its highly dynamic nature. A large amount of researches have proposed to address the issue. Some of them, with very promising results, adopt the Focused Time Lagged FeedForward Network (FTLFN), a type of Artificial Neural Network (ANN) that offers memory mechanisms capable of detecting temporal patterns. In many cases, however, the neural network’s capacities are not properly explored, being limited to tests with a minimum set of parameters. Besides, most of the studies on stock price prediction focus on low-frequency periods, such as days or months. On the other hand, due to the ease of access to information in the last decades and the automation of trades in stock market, these are getting more oftenly executed over short-term horizons, like hours, minutes or seconds. Therefore, there is a need to expand the knowledge related to forecasts in this scenario. With that in mind, this research has the objective of investigating the FTLFN’s potential on stock price forecasting over the intraday period. Its short-term memory and hidden layer size are widely and de eply explored, so the impact of different configurations on the accuracy results could be measured. Also, Technical Analysis indicators are built and utilized as input signals to the network, with their possible contributions to stock prediction being verified. From a general perspective, the work proposes the extention of the understanding regarding the FTLFN’s forecasting capabilities, as well as its use with intraday financial time series, which still require further exploration in literature. The obtained results show that, as human investors do, also FLTFNs are capable of taking enormous advantage from input signals’ history on providing better prediction quality within the proposed context. The same cannot be said for the supporting Technical Analysis indicators chosen, since they mostly increase forecasting errors. Evidences are presented based on the experimentation over several sets, bringing robustness to the conclusions and allowing the methodology and the results to serve as base for future researches related to predictions on high-frequency trading scenarios.

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2016-05-02T16:48:04Z No. of bitstreams: 1 William de Sousa Dias_.pdf: 766364 bytes, checksum: 83373201a1819852a99d5c675653eb18 (MD5) / Made available in DSpace on 2016-05-02T16:48:04Z (GMT). No. of bitstreams: 1 William de Sousa Dias_.pdf: 766364 bytes, checksum: 83373201a1819852a99d5c675653eb18 (MD5) Previous issue date: 2016-02-29 / Nenhuma / A tomada de decisão em investimentos tem representado um dos principais desafios com os quais se deparam os investidores ao longo do tempo. Constatam-se ainda muitas controvérsias quanto à capacidade das análises existentes em obter retornos positivos no mercado financeiro. Nesse sentido, o objetivo deste trabalho se concentra em avaliar o resultado financeiro da seleção de ações com o uso de dois critérios, um Fundamentalista e um Técnico, isolados e de forma combinada, em comparação ao retorno do índice de referência (Índice BOVESPA). Para a consecução desse propósito, primeiramente promoveu-se uma revisão teórica visando encontrar pontos de similaridades, conflitos e complementaridades dos dois tipos de análises, e como resultado, concluiu-se que a Análise Fundamentalista evidencia critérios melhor definidos para a escolha de qual ação comprar; já a Análise Técnica fornece ferramentas que definem com mais clareza o momento de comprar ou de vender determinada ação. A partir desse resultado, efetuou-se uma simulação computacional de três estratégias com dados fundamentalistas e técnicos do período de janeiro de 2010 a dezembro de 2014: uma Técnica, com o uso de médias móveis (aritmética e exponencial); uma Fundamentalista, utilizando-se a relação Book-to-Market (BM) como critério de seleção; e uma análise Híbrida, que aplicou o critério fundamentalista para a escolha de qual ação comprar e o critério técnico para decidir a hora de comprar ou de vender. Nas três simulações, as ações foram classificadas em grupos a partir de uma análise de Cluster, a qual adota o coeficiente de Spearman ao quadrado, de forma que duas ações do mesmo grupo não poderiam compor a carteira ao mesmo tempo. A partir da análise com o emprego do MaxiMax, do MaxiMin e do Minimax, e aplicando um teste estatístico de comparação por ranqueamento, depreendeu-se que a estratégia Fundamentalista desempenhou significantemente melhor em todos os cenários simulados (ao nível de 0,05), e que as abordagens Híbrida e Fundamentalista tiveram desempenho melhor que o IBOV. No entanto testou-se somente um tipo de cada abordagem. / Decision-making on investments has represented one of the main challenges faced by investors over time. There are still many controversies about the ability of existing analyzes to obtain positive returns in financial markets. In this sense, the objective of this work focus on evaluating the financial outcome of the stock selection using two criteria, one Fundamentalist and a Technical, isolated and in combination, compared to the return of the benchmark index (Bovespa Index). To achieve this purpose, first a theoretical review was promoted to find points of similarities, conflicts and complementarities of both types of analysis, and as a result, it was concluded that the Fundamentalist Analysis shows better defined criteria for choosing what share to buy; By the other hand Technical Analysis provides tools to define more clearly the time to buy or sell a particular share. From this result, a computer simulation of three strategies with fundamentalists and technical data for the period January 2010 to December 2014 was performed: A Technical with the use of moving averages (arithmetic and exponential); Fundamentalist one, using the ratio Book-to-Market (BM) as selection criteria; and a Hybrid analysis that applied the fundamental criterion for choosing which share to buy and technical criteria to decide the time to buy or sell. In all three simulations, the shares were classified in groups from a cluster analysis, which adopts the coefficient Spearman², so that two shares of the same group could not render the portfolio at the same time. From the Maximax, the Maximin and Minimax, and applying a statistical test compared by ranking was concluded that the Fundamentalist strategy played significantly better in all simulated scenarios (at 0.05). The Hybrid and Fundamentalist analysis was better than IBOV’s performance. However it was tested only one of each approach.

Este trabalho apresenta um método de predição não linear de séries temporais econômicas. O método baseia-se na análise técnica e fundamentalista de cotação de ações, filtragem wavelet, seleção de padrões e redes neurais artificiais. No modelo técnico emprega-se a transformada wavelet para filtrar a série temporal econômica de comportamentos aleatórios ou não econômicos. Após a filtragem dos dados o algoritmo de projeções sucessivas é utilizado para a seleção de padrões de treinamento para a rede neural artificial, com o objetivo de selecionar os padrões de comportamento mais importantes na série. No modelo fundamentalista utiliza-se variáveis econômicas que podem estar correlacionadas com a série, com o objetivo de aprimorar a predição da série na rede neural artificial. Para avaliação do método são utilizados dados de séries temporais econômicas referentes à cotação de preços de ações negociadas na bolsa de valores de São Paulo, onde os resultados da predição do comportamento futuro são comparados com modelos matemáticos clássicos e com o modelo convencional, que se baseia somente na análise técnica. Apresenta-se uma comparação dos resultados entre modelos técnicos, modelos matemáticos e o método proposto. O modelo matemático utilizado (ARIMA) apresentou seu melhor desempenho em séries com pouca variância, porém com desempenho inferior quando comparado com o modelo técnico e com o método proposto. A avaliação do erro de predição em termos de RMSEP evidenciou que o método proposto apresenta os melhores resultados em relação aos demais métodos. / This work presents a method for predicting nonlinear economic time series. The method is based on fundamental and technical analysis of script quotation, a multiscale wavelet filtering, pattern selection and artificial neural networks. In the technical model is used the wavelet transform in order to filter the economic time series from random or not economic behaviors. After the data filtering, the successive projections algorithm was used for the training pattern selection to the artificial neural network. In the fundamentalist model is used financial and macroeconomics variables that is correlated with the time serie in order to improve the network forecasting. For the evaluation of the proposed method are used temporal series data related to scrips prices quotation of São Paulo stock market. It presents a comparison of the results between technical model, mathematical model and proposed method. The mathematical model (ARIMA) presented better results in series with few variance, however have low performance when compared with the technical model and with the proposed method. The prediction error evaluation shows that the proposed method has better results than the other methods.

Med den nya omvärldsutvecklingen där NATO moderniserar sitt missilförsvar i Europa samt att Ryssland har placerat taktiska ballistiska robotar i Kaliningrad påverkade Sveriges behov till att anskaffa förmågan att bekämpa ballistiska robotar. Sverige har därför anskaffat Patriotsystemet, dock utan att tillföra ett radarsystem för att invisa ballistiska robotar till luftvärnsförbandet. Missilförsvar är ett väl utforskat område, allt från bekämpningsförlopp till hur en sensorkedja ska se ut. Forskningen tar däremot inte upp vilka förutsättningar ett nyanskaffat radarsystem behöver innefatta för att bidra till att invisa ballistiska robotar för luftvärnsförbanden. I uppsatsen genomfördes en modellering, teknisk analys, där teorin missilförsvar en kedja av event nyttjades för att härleda krav på radarprestanda. En analys om hur organisationerna idag nyttjar radarsystem och hur de tekniska och taktiska kraven påverkar organisationen genomfördes med konceptet militär nytta. Resultatet visar att organisationen där radarsystemet tillförs behövde kompletteras med ett sensorkompani och säkerhetsförband för att uppfylla kravet till invisning. Mot bakgrunden av det scenario som togs fram för undersökning visar den tekniska analysen att radarn vara fordonsburen samt ha en räckvidd på 500 km och en höjdtäckning på 50 km. Den måste även vara kompatibel med Patriotsystemet samt en sensorkedja för strategiskt partnerskap. / With the new developments in international affairs, where NATO modernizes its missile defense system in Europe and Russia has placed tactical ballistic missiles in Kaliningrad, Sweden needs to acquire the ability to combat ballistic missiles. Sweden has therefore acquired the Patriot system, however without adding an early warning radar for the Air defense against ballistic missile. Missile defense is a well-explored area, ranging from missile defense events to how a sensor chain should function. The research does not, however, discuss the abilities an acquired radar system needs to help guide ballistic missiles for Air defense units. In this essay, a modeling and a technical analysis based on the theory Missile defense a chain of events are used to conclude requirements for radar performance. An analysis of how the organizations use radar systems today and how the technical requirements affect the organizations was implemented with the concept Military utility. The result indicates the organization there the radar system will be implemented needs to be reinforced with a sensor company and a security unit to meet the requirement for guidance. The background of the scenario that has been developed for the analysis shows that the radar should be integrated to a vehicle, have a range of 500 kilometers and a height coverage of 50 kilometers. Battle management systems must also be compatible with the Patriot system and the sensor chain for strategic partnership.

Cette thèse proposée utilise les ondelettes de Haar à créer de nouveaux indicateurs techniques, d’en évaluer leurs performances afin de tester la validité de l’efficience faible des marchés financiers. L’approche choisie vise à mettre en œuvre les capacités des indicateurs techniques à capter la mémoire longue présente dans les indices boursiers américains et européens à travers l’estimation de la tendance par le processus de lissage. De plus, cette dernière est une composante importante dans les séries économiques et financières. En effet, elle a fait l’objet d’innombrables investigations tant en analyse technique, qu’en traitement du signal et dans la théorie des cycles économiques. Toutefois, sa présence n’entre pas en ligne de compte dans la théorie classique de la finance, car les principaux modèles utilisés se focalisent sur les variations des cours boursiers. À cet effet, la tendance constitue une source de non-stationnarité entraînant des difficultés majeures pour la modélisation économétrique ou financière. Exploiter cette tendance s’affranchit, dans ce cas, des hypothèses de non-stationnarité tendancielle ou de racine unitaire. En plus, à l’issue des résultats que nous avons obtenus à partir du modèle à changement de régime. Nous confirmons qu’il est possible d’exploiter la présence de mémoire longue dans les cours, et également de battre le marché en présence de coûts de transactions sur les marchés américains et européens. / This proposed thesis uses the Haar wavelets to create new technical indicators, to evaluate their performance in order to test the validity of the weak form of efficient market hypothesis. The chosen approach aims to implement the capabilities of technical indicators to capture the long memory present in the US and European stock indices through the estimation of the trend by the smoothing process. Moreover, the trend is an important component in the economic and financial series. Indeed, it has been the subject of innumerable investigations in technical analysis, in signal processing and in the theory business cycle theory. However, its presence is not taken into account in the classic theory of finance because the main models used focus on changes in stock prices. For this purpose, the trend constitutes a source of non-stationarity leading to major difficulties for econometric or financial modeling. Exploit trend is freed, in this case, from the hypotheses of tendancy or unit root. In addition, the issue of the results we obtained from the regime change model. We confirm that it is possible to exploit the presence of long memory in the series, and also to beat the market in the presence of transaction costs on the American and European markets.

[pt] A análise técnica ou grafismo consiste na identificação visual de padrões geométricos em gráficos de séries de preços de mercado com o objetivo de antecipar tendências de preço. Esta Dissertação revisita a questão da validação estatística da análise técnica, que tem sido estudada na literatura sem os devidos cuidados com os problemas de heterogeneidade e de dependência estatística dos dados analisados - agrupamento de séries de retornos referentes a diversos ativos financeiros distintos. O objetivo central deste estudo consiste em resolver o primeiro problema citado, através de uma metodologia para homogeneizar os ativos no que concerne às distribuições de probabilidades de suas séries de retorno. Os passos gerais desta metodologia envolvem a identificação dos processos estocásticos geradores dos retornos dos ativos, o agrupamento de ativos semelhantes e, finalmente, a análise de presença, ou ausência, de informação advinda dos padrões de preços. Como ilustração, são analisadas séries de diversos ativos do mercado financeiro mundial. A nossa investigação verifica a existência de conteúdo informativo estatisticamente significante em dois dos três padrões usualmente identificados na análise técnica, a saber: triângulos retângulos e head & shoulders. / [en] Technical analysis or charting aims on visually identifying geometrical patterns in price charts in order to anticipate price trends. This dissertation revisits the issue of technical analysis statistical validation, which has been tackled in the literature without taking care of the presence of heterogeneity and statistical dependence in the analyzed data - agglutinated return time series from many distinct securities. The main purpose of this study is to address the first cited problem by suggesting a methodology to homogenize the securities according to the probability distributions of their return series. The general steps of the methodology go through the identification of the data generating stochastic processes for the security returns, the clustering of similar securities and, finally, the analysis of the presence, or absence, of informational content coming from those price patterns. We illustrate the proposed methodology with several financial securities of the global market. Our investigation shows that there is a statistically significant informational content in two out of the three common patterns usually found through technical analysis, namely: triangles, rectangle and head & shoulders.

Since the late 1970s much interest has been shown in the development of electric guitar technique. Advances have been considerable, enabling players to explore new genres and repertoires but development methodologies have remained woefully fragmented. A new approach that sets out to promote electric guitar technique with development methodology is the purpose of this study. To this end, a process of transcription combined with an advanced technical analysis has been undertaken including a full categorization of the technical subgroups extant within each Caprice. The hypothesis behind this task has been to ascertain whether a ‘technical essence’ could be discovered in the Caprices and how that could be imparted in the process of transcription. Transcribing the 24 Caprices for the electric guitar disclosed the technical components required for development which were then reduced to their constitute elements. The virtuosity and variation within the Caprices ensured that the each identified technique was developed to a high degree. The subjective nature of transcription ensured that multiple solutions were explored when a single solution to a technical problem was not obvious. The analysis section of the study demonstrated that three fundamental techniques were required to play all 24 Caprices: alternate-picking, sweep-picking, and hammer-ons and pull-offs. The analyses also provided trends showing how each technique needed to be developed to comprehensively cover all twenty-four pieces. In conclusion, the hypothesis was found to be correct.

技術分析一直是許多研究的熱門主題，也被眾多市場參與者廣泛運用在市場交易，而最普遍且最受歡迎的技術分析工具即為移動平均法。 本研究設計三種移動平均交易方法（一種只考慮收盤價，一種考慮收盤價及交易量，而另一種則將交易量作為收盤價的權重），每種交易方法皆使用五天為短期移動平均天數，十天、五十天、一百天、一百五十天、兩百天為長期移動平均天數，總計十五種移動平均交易規則，運用在臺灣中型100成分股以產生買進與賣出訊號，並依訊號進行交易動作，進而在未考慮交易成本的假設下計算出單次交易的平均報酬、平均持有天數，以及Hit ratio（正報酬的交易次數占總交易次數的比例），藉以探討移動平均法在此種股票的獲利性。而以交易量為價格權重來產生移動平均交易方法是基於相信帶有較高交易量的價格較有意義，盼藉以測試此種方法是否正如預期，相較於一般傳統的價格移動平均法有更好的績效。 本研究雖然未考慮交易成本，但呈現的單次交易平均報酬可以提供讀者與實際臺灣股票市場交易成本作比較，藉以了解考慮交易成本後的情況。而本研究除了呈現所有成分股單次交易的平均報酬、平均持有天數及Hit ratio的平均值，也將成分股依照ICB行業分類指標分成幾個主要產業，並呈現各產業內成分股的平均值，企圖了解特定交易方法是否在特定產業有較好的績效。 結果顯示，產生最好績效的移動平均交易方法也僅能有一半的交易次數得到正報酬，而就整體而言，將交易量作為價格權重的移動平均方法，也沒有產生相較於傳統價格移動平均法更好的績效，因此可以說，這類的技術分析對於這些股票無法有較好的績效。 / Technical analysis has been widely studied and used by many researchers and market participants. The most common and popular technical trading rule is moving average since it is mathematically well defined and used by many analysts. This article examines the profitability of technical analysis for FTSE TWSE Mid-Cap Taiwan 100 Index constituents under the hypothesis of no transaction costs. It uses three strategies (Price Strategy, Price and Volume Strategy, and PV Strategy) and fifteen moving average rules to generate buy and sell signals, and then compute average returns per trading, average holding days per trading, and hit ratios to see the profitability. It is believed that prices come with high volumes are more meaningful than those with low volumes. All of these strategies and trading rules are not only used for all constituents of FTSE TWSE Mid-Cap Taiwan 100 Index without consid-ering industry classifications but also for each major industry classifications of these constituents. Therefore, we can understand whether specific trading rules have better performances for specific industries of these stocks. The results are not that optimistic. Overall Price and Volume Strategy has the best results of hit ratio, however, the highest value is barely 50%, which means it can only have a half trading times positive returns. As for PV Strategy which uses weighted price moving average to trade, the performance has no significantly better than using simple price moving average rule. It can say that Technical Analysis like moving average can hardly have good performances on these stocks.

學術上不接受技術分析等方法，認為股價已經在市場上充分反應，過去的歷史股價不能對未來進行預測。然而，業界或一般的投資人，卻往往把技術分析拿來做為買賣的依據。實際上以歷史資料做模擬交易，卻可以發現許多技術分析的法則在某些市場、股票、期間之中，可以獲得相對於買進賣出更好的報酬。有趣的是，任何一種操作法則或是特定一組參數選擇，在樣本外的操作則無法完全發現同樣的結果。故以技術分析所獲得的超額報酬，究竟是此機制有效還是單純運氣成分，許多技術分析的文獻以及著作往往著墨甚少。 本論文利用狀態轉換模型(Regime Switching Model)捕捉台灣加權股價指數，將股價的動態分為上漲以及下跌兩種狀態，並估計其市場參數—漲跌速度、漲跌速度標準差、轉換機率。其次將所估計的市場參數做為模擬的依據，可發現在單純隨機的環境下，某些市場參數組合存在移動平均線的交易策略明顯優於買進持有策略。研究中以敏感度分析的方法，呈現各個單一市場參數的改變情形，對於操作績效影響的方向。 最後將2001~2010的的台灣加權股價指數，估計市場參數並找尋當下最適的移動平均組合，允許每季重新調整參數，並實際以收盤價做為買賣模擬。結果發現移動平均線操作，確實能提供比買進持有更好的報酬，並減低每年報酬率變異。

This thesis further investigates the effectiveness of 15 variable moving average strategies that mimic the trading rules used in the study by Brock, Lakonishok, and LeBaron (1992). Instead of applying these strategies to developed markets, unique characteristics of emerging markets offer opportunity to investors that warrant further research. Before transaction costs, all 15 variable moving average strategies outperform the naïve benchmark strategy of buying and holding different emerging market ETF's over the volatile period of 858 trading days. However, the variable moving averages perform poorly in the "bubble" market cycle. In fact, sell signals become more unprofitable than buy signals are profitable. Furthermore, variations of 4 of 5 variable moving average strategies demonstrate significant prospects of returning consistent abnormal returns after adjusting for transaction costs and risk.