Die vorliegende Diplomarbeit beschäftigt sich mit zwei Aspekten der statistischen Eigenschaften von Clusterverfahren. Zum einen geht die Arbeit auf die Frage der Existenz von unterschiedlichen Clusteranalysemethoden zur Strukturfindung und deren unterschiedlichen Vorgehensweisen ein. Die Methode des Abstandes zwischen Mannigfaltigkeiten und die K-means Methode liefern ausgehend von gleichen Daten unterschiedliche Endclusterungen. Der zweite Teil dieser Arbeit beschäftigt sich näher mit den asymptotischen Eigenschaften des K-means Verfahrens. Hierbei ist die Menge der optimalen Clusterzentren konsistent. Bei Vergrößerung des Stichprobenumfangs gegen Unendlich konvergiert diese in Wahrscheinlichkeit gegen die Menge der Clusterzentren, die das Varianzkriterium minimiert. Ebenfalls konvergiert die Menge der optimalen Clusterzentren für n gegen Unendlich gegen eine Normalverteilung. Es hat sich dabei ergeben, dass die einzelnen Clusterzentren voneinander abhängen. / The following thesis describes two different views onto the statistical characterics of clustering procedures. At first it adresses the questions whether different clustering methods exist to ascertain the structure of clusters and in what ays the strategies of these methods differ from each other. The method of distance between the manifolds as well as the k-means method provide different final clusters based on equal initial data. The second part of the thesis concentrates on asymptotic properties of the k-means procedure. Here the amount of optimal clustering centres is consistent. If the size of the sample range is enlarged towards infinity, it also converges in probability towards the amount of clustering centres which minimized the whithin cluster sum of squares. Likewise the amount of optimal clustering centres converges for infinity towards the normal distribution. The main result shows that the individual clustering centres are dependent on each other.

<p><font face="F59" size="3"><font face="F59" size="3"> <p align="left">This study examines the relationship between stock returns and market segmentation. Monthly returns of stocks listed on the JSE from 1997-2007 are analysed using mostly the analytic factor and cluster analysis techniques. Evidence supporting the use of multi-index models in explaining the return generating process on the JSE is found. The results provide additional support for Van Rensburg (1997)'s hypothesis on market segmentation on the JSE.</p> </font></font></p>

The present thesis concerns middle-aged women’s subjective well-being (SWB). The interest is focused on the importance of childhood factors, social circumstances, and personality for middle-aged women’s general SWB. Data were taken from the longitudinal research program Individual Development and Adaptation (IDA, Magnusson &amp; Bergman, 2000) and concerned a sample of about 300 women. The main analyses were made on data collected at age 43, but data collected at age 13 and age 49 were also used to elucidate the purposes of this thesis. The results can be summarized as follows: 1) In a Swedish sample of middle-aged women, social circumstances had only a moderate effect on general SWB variables. The strongest relationship was found between marital status and global life satisfaction. When personality factors were controlled for, they wiped out nearly all relationships between the social circumstances variables and SWB, except for those between global life satisfaction and marital status or unemployment; 2) The level of general SWB was found to be considerably higher for Swedish employed women as compared to their counterparts in Lithuania and different socio-demographic variables predicted SWB in those two countries. For the Swedish sample, family-oriented variables were the strongest predictors of SWB, while for the Lithuanian sample income and educational level were more important; 3) Results from applying longitudinal structural equation modeling suggested that optimism in adolescence influenced optimism in middle age, which in its turn had both a direct influence on global life satisfaction and an indirect influence via the negative affect dimension. In relation to a number of different adjustment factors measured in adolescence it was found that optimism was the only factor that was constantly related to SWB 30 years later; 4) Typical patterns of general SWB were identified. Cluster analyses at age 43 and age 49 separately resulted in similar well-functioning six cluster solutions at both ages, indicating structural stability across six years. In addition to the typical high/low/average SWB clusters that could be to some degree expected from variable-oriented results, a cluster with intense affect and one with very low GLS emerged. All clusters except the latter one showed individual stability across six years.

The present research intended to examine the relationship between childhood stable peer status and adjustment in midadolescence for both genders, and adjustment in early and middle adulthood for women. One-year stably peer rejected, popular, and average boys and girls were identified by an age 10 and age 11 sociometric classification procedure using positive nominations and rank-ordering. These groups were examined in midadolescence. Findings indicated that stable peer rejection in childhood was associated with negative school adjustment and problems in peer relations in adolescence for both genders, and that many rejected boys were school dropouts. However, rejected children did not associate with deviant peers or show more antisocial behavior than their counterparts. For the adulthood follow-up, cluster analysis using LICUR was applied on rank-ordering data from age 10 and age 13 in order to identify 3-year stably rejected, popular, and average groups of girls. A methodological comparison with a standard sociometric method showed that the applied cluster analysis seems to be a useful additional tool in the arsenal of sociometric classification methods. Furthermore, one seems to arrive at larger stable peer status groups with this method than with other sociometric classification methods. The longitudinal follow-up into adulthood showed that rejected girls were at increased risk for criminal offending and alcohol abuse in young adulthood. Childhood aggressive behavior explained the significant association between peer rejection and criminality. There were no significant differences between the stable peer status clusters in the midadulthood adjustment areas of social relations, health, occupation, and subjective well-being, except that popular girls had achieved more successful vocational careers (which in its turn was explained by their higher academic achievement in childhood). Constraints and implications for future research were discussed.

A driving cycle is a velocity profile over time. Driving cycles can be used for environmental classification of cars and to evaluate vehicle performance. The benefit by using stochastic driving cycles instead of predefined driving cycles, i.e. the New European Driving Cycle, is for instance that the risk of cycle beating is reduced. Different methods to generate stochastic driving cycles based on real-world data have been used around the world, but the representativeness of the generated driving cycles has been difficult to ensure. The possibility to generate stochastic driving cycles that captures specific features from a set of real-world driving cycles is studied. Data from more than 500 real-world trips has been processed and categorized. The driving cycles are merged into several transition probability matrices (TPMs), where each element corresponds to a specific state defined by its velocity and acceleration. The TPMs are used with Markov chain theory to generate stochastic driving cycles. The driving cycles are validated using percentile limits on a set of characteristic variables, that are obtained from statistical analysis of real-world driving cycles. The distribution of the generated driving cycles is investigated and compared to real-world driving cycles distribution. The generated driving cycles proves to represent the original set of real-world driving cycles in terms of key variables determined through statistical analysis. Four different methods are used to determine which statistical variables that describes the features of the provided driving cycles. Two of the methods uses regression analysis. Hierarchical clustering of statistical variables is proposed as a third alternative, and the last method combines the cluster analysis with the regression analysis. The entire process is automated and a graphical user interface is developed in Matlab to facilitate the use of the software. / En körcykel är en beskriving av hur hastigheten för ett fordon ändras under en körning. Körcykler används bland annat till att miljöklassa bilar och för att utvärdera fordonsprestanda. Olika metoder för att generera stokastiska körcykler baserade på verklig data har använts runt om i världen, men det har varit svårt att efterlikna naturliga körcykler. Möjligheten att generera stokastiska körcykler som representerar en uppsättning naturliga körcykler studeras. Data från över 500 körcykler bearbetas och kategoriseras. Dessa används för att skapa överergångsmatriser där varje element motsvarar ett visst tillstånd, med hastighet och acceleration som tillståndsvariabler. Matrisen tillsammans med teorin om Markovkedjor används för att generera stokastiska körcykler. De genererade körcyklerna valideras med hjälp percentilgränser för ett antal karaktäristiska variabler som beräknats för de naturliga körcyklerna. Hastighets- och accelerationsfördelningen hos de genererade körcyklerna studeras och jämförs med de naturliga körcyklerna för att säkerställa att de är representativa. Statistiska egenskaper jämfördes och de genererade körcyklerna visade sig likna den ursprungliga uppsättningen körcykler. Fyra olika metoder används för att bestämma vilka statistiska variabler som beskriver de naturliga körcyklerna. Två av metoderna använder regressionsanalys. Hierarkisk klustring av statistiska variabler föreslås som ett tredje alternativ. Den sista metoden kombinerar klusteranalysen med regressionsanalysen. Hela processen är automatiserad och ett grafiskt användargränssnitt har utvecklats i Matlab för att underlätta användningen av programmet.

Background: Women and men in Sweden are in paid work to almost the same extent, but are found in different occupations and positions in the labour market. Still, women perform the bulk of the unpaid domestic work at home. Gendered inequalities in these respects leave women and men exposed to different work environments and responsibilities, which in turn can have gendered health consequences. In public health research there is a lack of studies on domestic work that include women and men, as well as a lack of qualitative studies exploring individuals’ experiences of domestic work and mental health. At the workplace level, few attempts have been made to analyse how several dimensions of gender equality at workplaces are related to health status and there is a lack of studies with a contextual approach combining many different variables that are at play simultaneously. Because of the cross-sectional design of previous studies on paid and domestic work there is a lack of analyses taking possible health-related selection into account, which makes it difficult to ascertain whether gender equality leads to better health or if good health is a prerequisite for gender equality. Aim: The aim of this thesis was to analyse gender relations of work (at workplaces and at home) in relation to mental illness among women and men. Methods: The thesis was based on data from the Northern Swedish Cohort. The baseline survey was conducted in 1981 when the participants were 16 years old (n= 1080, 574 boys and 506 girls), with follow-up at age 18, 21, 30 and 42. The response rate was 94 % throughout the last follow-up in 2007. Data from the Northern Swedish Cohort were supplemented with register data about the employees at the participants’ workplaces. The analysis methods for the questionnaire and register data were logistic regression analysis and cluster analysis. Interviews were performed with four women and four men in the Northern Swedish Cohort and were analysed with a Grounded Theory approach. Results: Women had overall greater responsibility for domestic work. Gender inequality in responsibility for domestic work and perceptions of gender inequality in the couple relationship (after adjustments for background variables and previous psychological distress) were associated with psychological distress among women and men. However, among men the relation between domestic work inequalities and psychological distress was affected by socioeconomic position relative to the partner. Having less responsibility for domestic work and a partner with higher socioeconomic position was associated with psychological distress among men. The qualitative analysis showed that gender relations were an important part of how the domestic work was unequally organised and related to experiences of mental illness among women and men. Among women the high burden of domestic work was experienced as an obstacle to experiencing good health. Among men the experience of being trapped in an outmoded masculinity was related to feelings of stress. At the workplace level, patterns of gender inequality were associated with psychological distress among women, but not among men. However, the most gender-equal pattern was related to lower as well as more similar levels of mental illness among women and men, which supports a convergence in health when women’s and men’s work conditions become more similar. Conclusion: Gender equality at home and at work is central for reducing mental illness among both women and men, but also for achieving a good average health status in the population, which is a central public health target. When investigating social inequalities in health, gender perspectives are of great importance for deepening the understanding of how and why gender inequalities in paid and domestic work are related to mental illness. Integrating gender perspectives into public health policy could be a way to acknowledge power relations that hinder good public health. / Bakgrund: Kvinnor och män i Sverige yrkesarbetar i nästan samma sträckning, men återfinns i olika yrken och positioner på arbetsmarknaden. Samtidigt utför kvinnor merparten av det obetalda arbetet i hemmet. Könade orättvisor i dessa avseenden innebär att kvinnor och män utsätts för olika arbetsmiljöer och ansvarsområden, vilket kan medföra könade hälsokonsekvenser. Dock är det få folkhälsovetenskapliga studier om obetalt arbete i hemmet och hälsa som inkluderar både kvinnor och män. Det är också en brist på kvalitativa studier som undersöker individers upplevelser av det obetalda arbetet i hemmet och psykisk hälsa. På arbetsplatsnivå har det gjorts få försök att analysera hur olika dimensioner av jämställdhet på arbetsplatser är relaterade till hälsostatus, och det är en brist på studier med en kontextuell metod som tar hänsyn till hur flera olika variabler samspelar. Tidigare folkhälsovetenskapliga studier som undersökt förvärvsarbete och hemarbete har framförallt använt sig av en tvärsnittsdesign, vilket gjort det svårt att utreda riktningen av samband mellan jämställdhet och psykisk hälsa, det vill säga om jämställdhet leder till bättre hälsa eller om god hälsa är en förutsättning för jämställdhet. Syfte: Syftet med denna avhandling var att analysera genusrelationer i arbete (på arbetsplatser och i hemmet) i förhållande till självskattad psykisk ohälsa bland kvinnor och män. Metod: Avhandlingen bygger på data från Luleåkohorten. Baslinjeundersökningen genomfördes år 1981 då deltagarna var 16 år (n=1080, 574 pojkar och 506 flickor). Uppföljningar har genomförts vid 18, 21, 30 och 42 års ålder och svarsfrekvensen var 94% under det senaste uppföljningen år 2007. Data från Luleåkohorten kompletterades med registerdata om anställda på deltagarnas arbetsplatser. Analysmetoder för enkät- och registerdata var logistisk regressionsanalys och klusteranalys. Intervjuer har genomförts med fyra kvinnor och fyra män i kohorten och analyserats med grundad teori. Resultat: Kvinnor hade totalt sett ett större och män ett mindre ansvar för det obetalda arbetet i hemmet. Upplevelse av bristande jämställdhet i ansvar för obetalt arbete i hemmet och i parrelationen hade ett samband med psykisk ohälsa bland kvinnor och män (även efter justering för bakgrundsvariabler och tidigare psykisk ohälsa). Att ha mindre ansvar för det obetalda arbetet i hemmet och en partner med högre socioekonomisk klass hade även ett samband med psykisk ohälsa bland män. Den kvalitativa analysen visade att genusrelationer var en viktig del i en ojämställd organisering av hemarbete och relaterade till upplevelser av psykisk ohälsa bland kvinnor och män. En hög belastning av hemarbete var ett hinder för kvinnor att uppleva en god hälsa. Upplevelser av att vara instängd i en omodern maskulinitet innebar en stressfull situation bland män. Mönster av ojämställdhet på arbetsplatser hade ett samband med psykisk ohälsa för kvinnor, men inte för män. Däremot visade sig de mest jämställda arbetsplats-mönstern vara relaterade till lägre och mer liknande nivåer av psykisk ohälsa bland kvinnor och män, vilket stödjer en konvergens i hälsa när kvinnor och mäns arbetsplats förhållanden är likvärdiga. Slutsats: Jämställdhet på arbetsplatser och i hemmet är viktigt för att minska psykisk ohälsa både bland kvinnor och män, men också för att uppnå en god genomsnittlig hälsa i befolkningen, något som är ett centralt folkhälsomål. Vid analyser av social ojämlikhet i hälsa är genusperspektiv av stor betydelse för att fördjupa förståelsen om hur och varför ojämställdhet i på arbetsplatser och i hemmet är relaterat till psykisk ohälsa. Att integrera genusperspektiv i folkhälsopolicy kan vara ett sätt att ta hänsyn till de maktrelationer som förhindrar en god folkhälsa.

Gendered practices of working life create gender inequalities through horizontal and vertical gender segregation in work, which may lead to gender inequalities in health experiences. The workplace is an important part of the social circumstances under which health opportunities and constraints are shaped. The workplace has also been identified as an important arena for gender constructions. Still, there is a lack of research about the relations between workplace gender equality and health experiences. The aim of this thesis was to explore gender equality and health experiences in a workplace setting. Qualitative and quantitative methods were used. In the qualitative studies all caregiving staff at two establishments providing care for elderly was invited to participate in focus groups (Papers I &amp; II). A moderator led 14 focusgroup discussions. Qualitative content analysis was used to analyse the transcribed discussions. For the quantitative studies questionnaire data from the Northern Swedish Cohort (n=836) were analysed and supplemented with register data about the participants’ workplaces. The register data were used to stratify the workplaces according to gender composition (paper IV) and to create gender equality indicators of the number of women and men at the workplace, education, salary and parental leave (Paper III). Cluster analysis was used to identify patterns of gender equality at the workplaces. Logistic regression analysis, adjusting for individual socio-demographics and previous psychological distress, were used to analyse psychological distress in relation to both clusters and gender compositions. This thesis identifies various workplace patterns of gender equality and how they are related to health experiences. The results from the focus group study showed that workplace stressors had a structural character, often originating from societal processes outside the own organization, whereas health resources had a relational character and were constructed within the organization (paper I). Gender equality was seen as a structural issue not connected to the individual health experiences and gender inequalities were justified through focusing on personalities and interests in work division (paper II). The cluster analysis resulted in six distinctive clusters with different workplace patterns of gender equality (paper III). The most gender-equal cluster was characterized by gender equality in salary and parental leave and was associated with the lowest prevalence of psychological distress, with no significant differences between women and men. The clusters were associated with psychological distress among women only. The highest odds for psychological distress among women were found in a traditional unequal cluster. Analyses of the gender composition at the workplace showed that the highest prevalence of psychological distress was found at workplaces with a mixed gender composition (paper IV). The psychosocial work environment was rather similar independent of the workplace gender composition.The factors most strongly associated with psychological distress were high demands and low control at workplaces with more men, being looked down upon at workplaces with a mixed gender composition, and social support at workplaces with more women. Gender perspectives highlight the importance of gender relations in research on work-related health. Gender inequalities at workplaces can be part of the explanation to women’s worse self-rated health. A multidimensional view of gender equality is necessary to understand health consequences of specific workplace situations. Workplaces are important arenas for health promotion activities and gender equality aspects needs to be taken into account to reach both women and men. Adequate health promotion needs to shift focus from individual health strategies to structural solutions that can challenge the root of the problem. / Ojämställdhet i arbetslivet kan innebära olika hälsokonsekvenser för kvinnor och män. Arbetsplatsen är en viktig del av de sociala omständigheter under vilka hälsomöjligheter och -begränsningar skapas. Arbetsplatsen har också identifierats som en viktig arena för genuskonstruktioner. Trots detta saknas forskning om relationerna mellan jämställdhet på arbetsplatser och hälsoupplevelser. Den här avhandlingen syftar till att studera sambanden mellan jämställdhet på arbetsplatser och de anställdas hälsoupplevelser. Avhandlingen använder både kvalitativa och kvantitativa metoder. I de kvalitativa studierna erbjöds all vårdpersonal på två äldreboenden att delta i fokusgruppsdiskussioner (Artikel I &amp; II). Sammanlagt genomfördes 14 fokusgrupper. Kvalitativ innehållsanalys användes för att analysera de transkriberade fokusgrupperna. I de kvantitativa studierna användes enkätdata från Luleåkohorten(n= 836) som kompletterats med registerdata om deltagarnas arbetsplatser. Registerdata användes för att stratifiera deltagarna utifrån könssammansättningen på deras arbetsplatser (artikel IV) och för att skapa jämställdhetsindikatorer för arbetsplatserna vad gäller andel män och kvinnor, utbildningsnivå, lön och föräldraledighet (artikel III). Klusteranalys användes för att identifiera mönster av jämställhet på arbetsplatserna. Skillnader i psykiska besvär mellan klustren respektive de olika könssammansättningarna analyserades med logistisk regressionsanalys som justerats för sociodemografiska variabler och tidigare psykiska besvär. Avhandlingen identifierar olika jämställdhetsmönster på arbetsplatsenoch hur de är relaterade till hälsoupplevelser. Resultaten från fokusgruppstudierna visade att de hälsorelaterade stressorerna hade en strukturell karaktär som ofta hade sitt ursprung i sociala processer utanför den egna organisationen. Hälsoresurserna var i hög utsträckning av relationell karaktär och skapades inom organisationen. Ojämställdhet sågs som ett strukturellt problem som inte kopplades till individuell hälsa. Ojämställdhet i arbetsfördelningen försvarades genom att fokusera på personligheter och individuella intressen. Kvantitativa analyser resulterade i sex kluster med olika jämställdhetsmönster på arbetsplatserna. Det mest jämställda klustret kännetecknades av arbetsplatser med jämställda löner och föräldraledighet och var relaterat till den lägsta förekomsten av psykiska besvär. I detta kluster hade också män och kvinnor samma förekomst av psykiska besvär. Bland kvinnorna fanns det skillnader mellan klusterna vad gäller psykiska besvär, men inte bland männen. Den högsta förekomsten av psykiska besvär bland kvinnor fanns i det traditionellt mest ojämställda klustret (artikel III) samt på arbetsplatser med en blandad könssammansättning (artikel IV). Variationen i den psykosociala arbetsmiljön var liten mellan arbetsplatser med olika könssammansättning men det fanns skillnader i sambanden mellan den psykosociala arbetsmiljön och psykiska besvär. De faktorer som hade starkast samband med psykiska besvär var höga krav och låg kontroll på arbetsplatser med en majoritet män, att bli ”sedd ner på” på könsblandade arbetsplatser, och socialt stöd på arbetsplatser med en majoritet kvinnor. Genusperspektiv är viktiga för att lyfta fram betydelsen av genusrelationer i forskning om arbetsrelaterad hälsa. Ojämställdhet på arbetsplatser kan vara en del av förklaringen till kvinnors sämre självskattade hälsa. En mångdimensionell syn på jämställdhet är nödvändig för att förstå hälsokonsekvenser av specifika arbetsplatssituationer. Arbetsplatser utgör också viktiga arenor för hälsofrämjande åtgärder och jämställdhetsaspekter måste beaktas för att dessa åtgärder ska nå både kvinnor och män på arbetsplatserna. Hälsofrämjande arbete måste skifta fokus från individuella, beteendeinriktade hälsostrategier till strukturella lösningar som kan utmana grunden till problemen.

In this study, a chlorinated-solvent contaminated groundwater site was used as the study site. Multivariate statistical analysis explains the huge and complicated current situation of the original data efficiently, concisely, and explicitly; it simplifies the original data into representative factors, or bases on the similarity between data to cluster and identify clustering outcome. The statistical software SPSS 12.0 was used to perform the multivariate statistical analysis to evaluate groundwater quality characteristics of this site. Results show that 20 analytical items of groundwater quality of the study site are simplified into seven major representative factors through factor analysis, including ¡§background¡¨, ¡§salt residual¡¨, ¡§hardness¡¨, ¡§ethylene chloride¡¨, ¡§alkalinity¡¨, ¡§organic pollutant¡¨, and ¡§chloroform¡¨. The factor score diagram was drawn according to the score of monitoring well on each factor and 89.6% of the variance could be obtained. This study used cluster analysis to cluster in two phrases, the groundwater quality monitoring wells were classified into seven clusters according to the similarity of monitored data nature and the differences between clusters. The groundwater quality characteristics and pollutant distributions of each cluster out this site were evaluated. The clustering result indicates that for the sixth cluster (where monitoring well SW-6 was the representative well), the average concentrations of chlorides such as 1,1-dichloroethylene, 1,1-dichloroethane, and cis-1,2-dichloroethylene were the highest among the clusters, indicating those the groundwater of nearby area might be polluted by chlorinated organic compounds. In addition, to evaluate whether the clustering of cluster analysis were appropriate or not, discriminant analysis is used to evaluate clustering accuracy, in which seven Fisher discriminant coefficient formulas that were exclusively suitable for this location were established. Then, the observed values were substituted to Fisher discriminant coefficient formula. Result shows that the monitoring well¡¦s clusters obtained from discriminant analysis were totally identical with the result of actual cluster analysis; the accuracy were 100%. After performing cross-validation analysis, the result shows that the accuracy were 80%, indicating the use of discriminant analysis (with forecasting function) to verify the clustering result of the cluster analysis was highly accurate. After analyzing the pollution condition of this site using time trend and space distribution, it were determined to conclude that trichloroethylene and 1,1-dichloroethylene were the major concerning pollutants; the pollutants appeared to be spreading on a large scale, so it was difficult to use the existing data to evaluate the pollution source. After assessing environmental medium characteristics and pollutant distribution of the site, this study suggests that the use of insitu bioremediation, which is cost-effective, can be applied as a remedial mothod.

The aim of this study is to investigate the distributions of polycyclic aromatic hydrocarbons (PAHs) in sediment and water (suspended particles and dissolved phase) samples collected from Dapeng Bay(coast¡Bbay¡Bwetland and mangrove). Chemical fingerprinting techniques and statistical analysis were applied to delineate the possible sources of PAHs in sediment and water samples. The average concentrations of total PAHs (£U51PAHs) ranged from 143 to 1595 ng/g dry weight (dw) in the sediments. Total PAH (£U46PAHs) concentrations varied from 11 to 38 ng/L in dissolved phase and from 1.4 to 15ng/L (252 to 772 ng/g dw) in suspended particulate phase, respectively. Comparing with sediment quality guidelines (SQGs), the total PAH concentrations in all sediment samples were below the criteria, suggesting that no adverse biological effects would arise from the PAHs in these sediments. The partition coefficients¡]KOC¡^values of PAHs were higher than predicted values. It might be attributed to soot particles which have extremely high sorption capacities. Total organic carbon¡]POC¡^and black carbon (BC) were significantly correlated with combustion PAH concentrations in Dapeng Bay. Combustion PAHs (Flt¡ÏPyr¡ÏBaA¡ÏChr¡ÏBbF¡ÏBkFa¡ÏBeP¡ÏBaP¡ÏIP¡ÏBghiP) concentrations were significantly correlated with black carbon (BC) in mangroves sediment cores and wetland A. In addition, total PAH concentrations were significantly correlated with combustion PAHs concentrations in dissolved phase. In suspended particulate phase, total PAH concentrations and combustion PAHs concentrations were significantly correlated with suspended solid (SS). Results from diagnostic ratios and hierarchical cluster analysis (HCA) both indicate that PAHs in the coast of Dapeng Bay were mainly from petrogenic sources, but from petrogenic and pyrogenic mixed sources in the wetland and mangrove inside the Dapeng Bay. In addition, the origin of perylene in Dapeng Bay and coastal sediments was mainly from the biogenic and pyrogenic, while it was from biogenic origin in wetland A and wetland B. However, the origin of perylene in mangrove M1 and mangrove M2 tended to be diagenetic.

In this paper, we use a Fama-French model and Markov regime-switching model to capture time series behavior of many financial variable. Alternatively, classification by cluster analysis help to learn the different characteristics of the sample between stock returns and risk factors. This empirical result shows that the excess return in the low volatility state tends to be greater than that in the high volatility state. The stock returns in each regime have a higher probability of remaining in their original state, especilly in low volatility state. This article also found the influence of risk factors affecting the stock returns is not symmetrical. In the state of low volatility, market factors and momentum effect have a significant influence in stock returns, and in the high volatility state, except the size effect, market and behavior factors have a significant influence in stock returns. Markov-switching models have proved to be useful for modeling a range of economic time series in the stock market. The regime-switching model has a superior performance in capturing the risk sensitivities of the stock return beyond the findings based on the Fama-French models. At last, we find the cluster analysis is feasible for the multi-factor model. The returns of mature companies have a primarily impact of market risk premium, while the major factor affecting returns with characteristics of growth companies is a investor sentiment. In addition, it is found that small companies¡¦ returns are vulnerable to investors sentiment. In this case, investors will invest based on stock's past performance, so the momentum effect significantly affect the stock returns.