Stochastic approximation methods in stochastic optimization

Stochastic approximation methods for stochastic optimization are considered. Reviewed the main methods of stochastic approximation: stochastic quasi-gradient algorithm, Kiefer-Wolfowitz algorithm and adaptive rules for them, simultaneous perturbation stochastic approximation (SPSA) algorithm. Suggested the model and the solution of the retailer's profit optimization problem and considered an application of the SQG-algorithm for the optimization problems with objective functions given in the form of ordinary differential equation.

Risk-return trade-off and autocorrelation

A trade-off between return and risk plays a central role in financial economics. The intertemporal capital asset pricing model (ICAPM) proposed by Merton (1973) provides a neoclassical theory for expected returns on risky assets. The model assumes that risk-averse investors (seeking to maximize their expected utility of lifetime consumption) demand compensation for bearing systematic market risk and the risk of unfavorable shifts in the investment opportunity set. Although the ICAPM postulates a positive relation between the conditional expected market return and its conditional variance, the empirical evidence on the sign of the risk-return trade-off is conflicting. In contrast, autocorrelation in stock returns is one of the most consistent and robust findings in empirical finance. While autocorrelation is often interpreted as a violation of market efficiency, it can also reflect factors such as market microstructure or time-varying risk premia. This doctoral thesis investigates a relation between the mixed risk-return trade-off results and autocorrelation in stock returns. The results suggest that, in the case of the US stock market, the relative contribution of the risk-return trade-off and autocorrelation in explaining the aggregate return fluctuates with volatility. This effect is then shown to be even more pronounced in the case of emerging stock markets. During high-volatility periods, expected returns can be described using rational (intertemporal) investors acting to maximize their expected utility. During lowvolatility periods, market-wide persistence in returns increases, leading to a failure of traditional equilibrium-model descriptions for expected returns. Consistent with this finding, traditional models yield conflicting evidence concerning the sign of the risk-return trade-off. The changing relevance of the risk-return trade-off and autocorrelation can be explained by heterogeneous agents or, more generally, by the inadequacy of the neoclassical view on asset pricing with unboundedly rational investors and perfect market efficiency. In the latter case, the empirical results imply that the neoclassical view is valid only under certain market conditions. This offers an economic explanation as to why it has been so difficult to detect a positive tradeoff between the conditional mean and variance of the aggregate stock return. The results highlight the importance, especially in the case of emerging stock markets, of noting both the risk-return trade-off and autocorrelation in applications that require estimates for expected returns.

Context-specific independence in graphical models

The theme of this thesis is context-speci c independence in graphical models. Considering a system of stochastic variables it is often the case that the variables are dependent of each other. This can, for instance, be seen by measuring the covariance between a pair of variables. Using graphical models, it is possible to visualize the dependence structure found in a set of stochastic variables. Using ordinary graphical models, such as Markov networks, Bayesian networks, and Gaussian graphical models, the type of dependencies that can be modeled is limited to marginal and conditional (in)dependencies. The models introduced in this thesis enable the graphical representation of context-speci c independencies, i.e. conditional independencies that hold only in a subset of the outcome space of the conditioning variables. In the articles included in this thesis, we introduce several types of graphical models that can represent context-speci c independencies. Models for both discrete variables and continuous variables are considered. A wide range of properties are examined for the introduced models, including identi ability, robustness, scoring, and optimization. In one article, a predictive classi er which utilizes context-speci c independence models is introduced. This classi er clearly demonstrates the potential bene ts of the introduced models. The purpose of the material included in the thesis prior to the articles is to provide the basic theory needed to understand the articles.

Novel genes and regulatory systems in epididymal differentiation and sperm maturation of the mouse.

Mammalian spermatozoa gain their fertilizing ability during maturation in the epididymis. Proteins and lipids secreted into the epididymal lumen remodel the sperm membrane, thereby providing the structure necessary for progressive motility and oocyte interaction. In the current study, genetically modified mouse models were utilized to determine the role of novel genes and regulatory systems in the postnatal development and function of the epididymis. Ablation of the mouse β-defensin, Defb41, altered the flagellar movements of sperm and reduced the ability of sperm to bind to the oocyte in vitro. The Defb41-deficient iCre knock-in mouse model was furthermore utilized to generate Dicer1 conditional knock-out (cKO) mice. DICER1 is required for production of mature microRNAs in the regulation of gene expression by RNA interference. Dicer1 cKO gave rise to dedifferentiation of the epididymal epithelium and an altered expression of genes involved in lipid synthesis. As a consequence, the cholesterol:polyunsaturated fatty acid ratio of the Dicer1 cKO sperm membrane was increased, which resulted in membrane instability and infertility. In conclusion, the results of the Defb41 study further support the important role of β-defensin family members in sperm maturation. The regulatory role of Dicer1 was also shown to be required for epididymal development. In addition, the study is the first to show a clear connection between lipid homeostasis in the epididymis and sperm membrane integrity. Taken together, the results give important new evidence on the regulatory system guiding epididymal development and function

Prenatal and perinatal risk factors for bipolar disorder

Bipolar disorder (BPD) is a severe mental disorder associated with considerable morbidity and mortality. Prenatal insults have been shown to be associated with later development of mental disorders and there is a growing interest in the potential role of prenatal and perinatal risk factors in the development of BPD. The aims of this thesis were to describe the overall study design of the Finnish Prenatal Study of Bipolar Disorders (FIPS-B) and demographic characteristics of the sample. Furthermore, it was aimed to examine the association of parental age, parental age difference, perinatal complications and maternal smoking during pregnancy with BPD. This thesis is based on FIPS-B, a nested case-control study using several nationwide registers. The cases included all people born in Finland between January 1st 1983 and December 31st 1998 and diagnosed with BPD according to the Finnish Hospital Discharge Register (FHDR) before December 31st 2008. Controls for this study were people who were without BPD, schizophrenia or diagnoses related to these disorders, identified from the Population Register Centre (PRC), and matched two-fold to the cases on sex, date of birth (+/- 30 days), and residence in Finland on the first day of diagnosis of the matched case. Conditional logistic regression models were used to examine the association between risk factors and BPD. This study included 1887 BPD cases and 3774 matched controls. The mean age at diagnosis was 19.3 years and females accounted for 68% of the cases. Mothers with the lowest educational level had the highest odds of having BPD in offspring. Being born in Eastern and Southern region of Finland increased the odds of having BPD later in life. A U-shaped distribution of odds ratio was observed between paternal age and BPD in the unadjusted analysis. Maternal age and parental age difference was not associated with BPD. Birth by planned caesarean section was associated with increased odd of BPD. Smoking during pregnancy was not associated with BPD in the adjusted analyses. Region of birth and maternal educational level were associated with BPD. Both young and old father’s age was associated with BPD. Most perinatal complications and maternal smoking during pregnancy were not associated with BPD. The findings of this thesis, considered together with previous literature, suggest that the pre- and perinatal risk factor profile varies among different psychiatric disorders.

Sentence Compression with Supervised Machine Learning

The subject of the thesis is automatic sentence compression with machine learning, so that the compressed sentences remain both grammatical and retain their essential meaning. There are multiple possible uses for the compression of natural language sentences. In this thesis the focus is generation of television program subtitles, which often are compressed version of the original script of the program. The main part of the thesis consists of machine learning experiments for automatic sentence compression using different approaches to the problem. The machine learning methods used for this work are linear-chain conditional random fields and support vector machines. Also we take a look which automatic text analysis methods provide useful features for the task. The data used for machine learning is supplied by Lingsoft Inc. and consists of subtitles in both compressed an uncompressed form. The models are compared to a baseline system and comparisons are made both automatically and also using human evaluation, because of the potentially subjective nature of the output. The best result is achieved using a CRF - sequence classification using a rich feature set. All text analysis methods help classification and most useful method is morphological analysis. Tutkielman aihe on suomenkielisten lauseiden automaattinen tiivistäminen koneellisesti, niin että lyhennetyt lauseet säilyttävät olennaisen informaationsa ja pysyvät kieliopillisina. Luonnollisen kielen lauseiden tiivistämiselle on monta käyttötarkoitusta, mutta tässä tutkielmassa aihetta lähestytään television ohjelmien tekstittämisen kautta, johon käytännössä kuuluu alkuperäisen tekstin lyhentäminen televisioruudulle paremmin sopivaksi. Tutkielmassa kokeillaan erilaisia koneoppimismenetelmiä tekstin automaatiseen lyhentämiseen ja tarkastellaan miten hyvin erilaiset luonnollisen kielen analyysimenetelmät tuottavat informaatiota, joka auttaa näitä menetelmiä lyhentämään lauseita. Lisäksi tarkastellaan minkälainen lähestymistapa tuottaa parhaan lopputuloksen. Käytetyt koneoppimismenetelmät ovat tukivektorikone ja lineaarisen sekvenssin mallinen CRF. Koneoppimisen tukena käytetään tekstityksiä niiden eri käsittelyvaiheissa, jotka on saatu Lingsoft OY:ltä. Luotuja malleja vertaillaan Lopulta mallien lopputuloksia evaluoidaan automaattisesti ja koska teksti lopputuksena on jossain määrin subjektiivinen myös ihmisarviointiin perustuen. Vertailukohtana toimii kirjallisuudesta poimittu menetelmä. Tutkielman tuloksena paras lopputulos saadaan aikaan käyttäen CRF sekvenssi-luokittelijaa laajalla piirrejoukolla. Kaikki kokeillut teksin analyysimenetelmät auttavat luokittelussa, joista tärkeimmän panoksen antaa morfologinen analyysi.

Variance reduction properties of dynamic hedging models - empirical evidence from BRICS countries

This Master’s Thesis analyses the effectiveness of different hedging models on BRICS (Brazil, Russia, India, China, and South Africa) countries. Hedging performance is examined by comparing two different dynamic hedging models to conventional OLS regression based model. The dynamic hedging models being employed are Constant Conditional Correlation (CCC) GARCH(1,1) and Dynamic Conditional Correlation (DCC) GARCH(1,1) with Student’s t-distribution. In order to capture the period of both Great Moderation and the latest financial crisis, the sample period extends from 2003 to 2014. To determine whether dynamic models outperform the conventional one, the reduction of portfolio variance for in-sample data with contemporaneous hedge ratios is first determined and then the holding period of the portfolios is extended to one and two days. In addition, the accuracy of hedge ratio forecasts is examined on the basis of out-of-sample variance reduction. The results are mixed and suggest that dynamic hedging models may not provide enough benefits to justify harder estimation and daily portfolio adjustment. In this sense, the results are consistent with the existing literature.

Economic capital allocation

For my Licentiate thesis, I conducted research on risk measures. Continuing with this research, I now focus on capital allocation. In the proportional capital allocation principle, the choice of risk measure plays a very important part. In the chapters Introduction and Basic concepts, we introduce three definitions of economic capital, discuss the purpose of capital allocation, give different viewpoints of capital allocation and present an overview of relevant literature. Risk measures are defined and the concept of coherent risk measure is introduced. Examples of important risk measures are given, e. g., Value at Risk (VaR), Tail Value at Risk (TVaR). We also discuss the implications of dependence and review some important distributions. In the following chapter on Capital allocation we introduce different principles for allocating capital. We prefer to work with the proportional allocation method. In the following chapter, Capital allocation based on tails, we focus on insurance business lines with heavy-tailed loss distribution. To emphasize capital allocation based on tails, we define the following risk measures: Conditional Expectation, Upper Tail Covariance and Tail Covariance Premium Adjusted (TCPA). In the final chapter, called Illustrative case study, we simulate two sets of data with five insurance business lines using Normal copulas and Cauchy copulas. The proportional capital allocation is calculated using TCPA as risk measure. It is compared with the result when VaR is used as risk measure and with covariance capital allocation. In this thesis, it is emphasized that no single allocation principle is perfect for all purposes. When focusing on the tail of losses, the allocation based on TCPA is a good one, since TCPA in a sense includes features of TVaR and Tail covariance.

Liquidity Risk in the Finnish Stock Market

This study explores the pricing of liquidity risk and its effect on stock returns in the Finnish stock market. In addition to that, it investigates whether there is a trend in liquidity risk. Finally, it analyzes whether the two chosen liquidity measures provide different results. The data consists of all the common shares listed in the Finnish stock market during the period of 1/1997–7/2015. To examine whether liquidity risk affects stock returns in the Finnish stock market, this study utilizes a conditional version of liquidity-adjusted capital asset pricing model (LCAPM) by Acharya and Pedersen (2005). Two recently proposed illiquidity measures – PQS and AdjILLIQ – are used in the empirical estimation to see whether there are differences in the results between the measures. The time-varying conditional liquidity risks are estimated by using a multivariate DCC-GARCH model, while the pricing of the liquidity risk is conducted by applying fixed effect panel regression. The results imply that investors in the Finnish stock market are willing to pay a premium to hedge from wealth shocks and having liquid assets during the declined market liquidity. However, investors are not willing to pay a premium for stocks with higher returns during illiquid markets. The total annualized illiquidity premiums found in the Finnish stock market are 1.77% and 1.04%, based on the PQS and AdjILLIQ measures, respectively. The study also shows that liquidity risk does not exhibit decreasing trend, and investors should consider liquidity risk in their portfolio diversification in the Finnish stock market.

Suomen taloussuhdanteiden kansainvälinen yhteisvaihtelu

Taloussuhdanteiden yhteisvaihtelun tutkimus on eräs taloustieteiden vanhimmista tutkimusaloista. Finanssikriisi ja euroalueen kohtaamat talousvaikeudet ovat kuitenkin nostaneet aiheen jälleen hyvin ajankohtaiseksi. Kuluneiden kahdenkymmenen vuoden aikana tutkimusalueesta on muodostunut erittäin laaja lukuisine näkökulmineen ja debatteineen. Tutkielman aiheena on Suomen taloussuhdanteiden kansainvälinen yhteisvaihtelu valittujen vertailumaiden kanssa. Vertailumaat ovat Ruotsi, Norja, Tanska, Saksa, Ranska, Iso-Britannia ja Yhdysvallat. Tutkielmaan valitut taloussuhdannetta kuvaavat muuttujat ovat reaalinen bruttokansantuote, yksityinen kokonaiskulutus ja teollisuustuotantoindeksi. Aineisto on kerätty Lappeenrannan tiedekirjaston Nelli-portaalin OECD iLibrary-tietokannasta ja se kattaa aikajakson 1960 Q1- 2014 Q4. Maakohtainen taloussuhdanne operationalisoidaan laskemalla ensimmäinen logaritminen differenssi, joka edustaa perinteistä reaalisuhdanneteoreettisen koulukunnan näkemystä taloussuhdanteesta. Tutkielman näkökulmaksi valitaan yhden maan näkökulma, joka on hieman harvinaisempi näkökulma verrattuna laajempiin alueellisiin näkökulmiin. Tutkimusmenetelminä käytetään Pearsonin korrelaatiokerrointa, Engle-Granger- sekä Johansenin yhteisintegroituvuustestejä ja VAR-GARCH-BEKK –mallilla laskettua dynaamista korrelaatiota, jotka lasketaan Suomen ja vertailumaiden välille maapareittain. Tuloksia tulkitaan suomalaisen vientiä vertailumaihin suunnittelevan yrityksen näkökulmasta. Tutkielman tulosten perusteella Engle-Grangerin menetelmällä laskettu samanaikainen yhteisintegroituvuus Suomen ja vertailumaiden välillä on epätodennäköistä. Kun yhteisintegroituvuuden annetaan riippua myös viiveistä, saadaan Johansenin menetelmällä yhteisintegroituvuus Suomen ja Yhdysvaltojen välille reaalisessa bruttokansantuotteessa, Suomen ja Saksan, Suomen ja Ranskan sekä Suomen ja Yhdysvaltojen välille yksityisessä kokonaiskulutuksessa sekä Suomen ja Norjan välille teollisuustuotantoindeksissä. Tulosten tulkintaa vaikeuttavat niiden malliriippuvuus ja informaatiokriteerien toisistaan poikkeavat mallisuositukset, joten yhteisintegroituvuus on mahdollinen myös muiden maaparien kohdalla. Dynaamisten korrelaatiokuvaajien perusteella maaparien välisen yhteisvaihtelun voimakkuus muuttuu ajan mukana. Finanssikriisin aikana kokonaistuotannossa on havaittavissa korkeampi korrelaatio, mutta korrelaatio palaa sen jälkeen perustasolleen. Kokonaiskulutuksen korrelaatio on kokonaistuotantoa alhaisempi ja pitemmissä aikajaksoissa vaihtelevaa.

Globalization of pharmaceutical R&D. Pharmaceutical R&D governance forms in People's Republic of China

Yrityksen selviytyminen ja menestyminen ovat riippuvaisia sen kyvystä innovoida, luoda tietoa ja hyödyntää tietämystä ja keksintöjä (Dunk ja Kilgore 2001). Yrityksen menestyminen erityisesti korkean teknologian alalla on siten suoraan riippuvainen sen T&T:stä, johon tehdyt investoinnit tuovat merkittäviä taloudellisia etuja yritykselle uusien tuotteiden, palveluiden ja prosessien muodossa (McEvily ja Chakravarthy 1999). Teknologinen etumatka ja sen tuotteistaminen innovatiivisiksi tarjoamiksi mahdollistaa monopolististen etujen saavuttamisen yrityksen kansainvälisessä kilpailussa (Lall 1977). Tämä kaltainen kilpailuetu voidaan saavuttaa yrityksen kyvyllä yhdistää maantieteellisesti hajautettu T&T:nsä tehokkaaksi verkostoksi (Porter 1986). Boehen (2008) mukaan T&T:n globalisoitumista voidaan johtaa eri hallintömuodoilla: T&T:n kansainvälistymisellä, T&T:n ulkomaille sijoittamisella ja T&T ulkomaille ulkoistamisella. T&T:n globalisoituminen on osa 2000-luvun taloudellista muutosta, ja sille on esitetty useita vaikuttavia tekijöitä, kuten kustannuserot, työvoimaresurssit, erityisosaamiskeskukset, paikallinen teknologia osaaminen ja kohdemarkkinoiden potentiaali (bardhan 2006; Norwood, ym. 2006; von Zedtwitz ja Gassmann 2002). Tutkimuksen on osoitettu eroavan tuotekehityksestä ja eri tekijöiden on osoitettu vaikuttavan niihin (von Zedtwitz ja Gassmann 2002; Leifer ja Triscari 1987). Samoin T&T on osoitettu olevan jatkumo perustavanlaatuisesta soveltavaan ja lääkekehityksen muodostavan vastaavan T&T jatkumon (Lall 1980; Iansiti 1993), jonka yksittäiset osat vaikuttavat sen hallintomuotoon. Tutkimus esittää eri tekijöitä voivan hyödyntää hallintomuodosta riippuen. Tätä tutkimusta varten tutkija haastatteli lääketeollisuuden johtajia Kiinassa vahvistaakseen tai hylätäkseen eri tekijöitä ja niiden suhdetta lääketeollisen T&T:n hallintomuotoihin. Markkinoiden todettiin olevan ensisijainen tekijä mutta myös kustannuserojen, insentiivien, työvoimaresurssien ja erityisosaamiskeskusten merkitys T&T:n globalisoitumiseen vaikuttavina tekijöinä vahvistettiin yhdessä perusvaatimusten ja riskitekijöiden kanssa. Tutkimuksessa vahvistetaan myös lääketeollisen T&T-jatkumon vaikutus ja esitetään viitekehys hallintomuodoille.

Globalization of pharmaceutical R&D Pharmaceutical R&D governance forms in People's Republic of China

Firm's survival and success, which are dependent on its ability to innovate, to create knowledge and to capitalize on inventions and know-how, is in essence directly linked to its R&D process (Dunk and Kilgore 2001). Especially in technology driven industries, such as the pharmaceuticals, there are significant positive returns to R&D investments through introduction of new or improved products and services (McEvily and Chakravarthy 1999). Technological lead and its transformation to innovative products as fruits of corporate R&D can be seen as monopolistic advantage that helps enterprises to compete in today’s market (Lall 1977). This competitive advantage can be derived from corporation's ability to integrate its activities across geographic locations (Porter 1986). According to Boehe (2008) globalization of R&D can executed with different governance forms: R&D internationalization, R&D offshoring or R&D offshore outsourcing. Globalization of R&D is intervened with the changes in global economy of the 21st century. Some studies argue for its influencing factors to be access to vast skilled labor pools and centers of excellence (Bardhan 2006). Other studies indicate the R&D cost differentials between countries to be the major expected benefit (Norwood et al. 2006). Von Zedtwitz and Gassmann (2002) presented benefits as divided to accessing markets and customers or to accessing local science and technology. This study proposes that based on governance form distinct factor derived benefits can be capitalized. To corroborate or refute factors and their relations on R&D globalization governance forms, an empirical study based on expert interviews of pharmaceutical directors was conducted in the People's Republic of China. The market was found to be the major influencing factor. Local requirements and adaptation were corroborated as factors connected with markets. Furthermore, influencing factors, such as labour, centers of excellence, cost, financial incentives were corroborated together with conditional and risk factors. Furthermore this research argues that the globalization of pharmaceutical R&D is dependent on the financial, scientific and operational requirements of the drug discovery stage. And thus establishes the influence of drug discovery's stages continuum on pharmaceutical R&D globalization. Finally, a R&D globalization governance form decision framework is proposed based on the frameworks presented in literature and author's corroborated empirical findings.