| In the field of investment,the most important is to study how to control a certain risk of buying and selling.Usually,we do not know the next tranche of trading is the profit or loss,but through certain trading strategies,we can while maintaining a certain expected return at the same time,maximize reduce risk encountered in the transaction process.At present,the statistical analysis method has been widely used in market analysis.Most investors tend to use the sample data to determine whether the future of the market is higher or lower.In the stock market,investors usually use profits,sales and debt to judge the fair value of the company,and through the horizontal comparison to select the better performance of the company.In commodity markets,investors are judged by the trend of supply and demand.Another popular analytical method is technical analysis.Technical analysis does not attempt to predict the market move through basic data,by contrast,technical analysis suggests that a market player with sufficient information does not have more information than the other participants.Therefore,technical analysts think that the price is the final performance of all the information after interaction.Therefore,technical analysis tends to market price as the starting point,through the mining market operation mechanism and then predict the price trend,so as to achieve the investment income.At the same time,in the field of investment there is a way of investment,it is not technology analysis that excessive reliance on personal,but according to some fixed down special rules,it produced different effects because of the use of different strategies in the process does not use,this method in a certain extent inhibited the investor’s subjective excessive investment behavior,which makes the strategy in the implementation process is more stable,and has a good predictability.This is the so-called quantitative investment technology.Quantitative trading process can be considered as a process of transaction process.Quantitative investment strategy is a combination of a series of technical analysis and statistical analysis,these analyses use price indicators or the use of complex indicators based on market price as a signal to buy or sell.Once a trading strategy is formed,we can use historical data to test it,to determine whether this strategy is established in history.Once established,it is likely to remain applicable in the future market.Again through the detection of historical data,we can use the strategy in different markets to simulate the transaction,through the simulated trading in the market,you can through the simulation results of dynamic programming,find a stable method to reduce transaction risk expected return unchanged conditions.In recent years,with the development of computer technology and the deepening of the development of the capital market,the quantitative investment model gradually presented a multi field,cross development trend.The research and application of the model has not only limited in the previous single investment logic and mathematical model of the derivation and evolution,accordingly,more and more many other areas of research is applied to the field of quantitative investment.The trend of the future development of quantitative models,to invest in logic based,was established on the basis of the investment logic required tools,the in each field of research,as long as with investment logic Xiang Wenhe.Using this logic of the establishment of,and then forming model.Because the quantitative investment model has a large number of production conditions,and the different stages of the investment model are different,so it is difficult to find a unified framework to study the similarities and differences of quantitative investment model.The purpose of this paper is to study the adaptability of the financial data and financial model,and find out that the financial data can reflect the inherent law of the target market in different capital markets.And in the study on the adaptability of model based on,try to establish a more effective quantitative model system and achieve better effect of stock,making in maintaining the expected return at the same time,maximize reduce transaction risk.In the research content,quantitative stock selection model this paper focuses on the research of stock investment behavior in the adaptation stage.Study found that the ultimate measure to quantify the different stock selection model adjustability is corresponding to the stock selection models can be screened with good discrimination of the stock portfolio.Firstly,this paper discusses the characteristics and the nature of the financial data in the capital market,and then discusses the establishment,processing and analysis of the financial data.The second step,to our country stock market data based,for different industries,the fitness of quantitative investment model under the different market style were studied,analyzes the advantages and disadvantages of many domestic and foreign quantitative stock selection model,and on this basis,establish quantitative investment model adaptability evaluation system,used to quantitative evaluation of investment model in different market adaptation degree.Through the different quantization model in China’s market adaptability comparison analysis,we found that: a quantitative model in different market adjustability of the key factor lies in the following two aspects: one is how to the stock market in the characterization of vector of validity analysis,that there is a stronger relationship screening characterizing factors and stock returns.The second is how to establish a stable and accurate mapping model,makes the internal relationship between the two can be more accurately to quantify,and effectively stock classification,with maximum discrimination of the stock portfolio.The third step,in the study on the adaptability of quantitative investment model based on,this paper studies the vector for the characterization of the stock market reaction of investment and income,has carried on the analysis to the stock market characterization vector,established quantitative investment characterization vector validity analysis model,from relativity,forecasting and sustainability of different dimension to measure the effectiveness and stock market characterization vector in a specific target market,to screen out reliable to vector for the characterization of the reaction of stock returns.The fourth step,after a quantitative investment characterization vector validity analysis model,we further study stock data and quantitative investment model,try to build a more effective and suitable for quantitative investment model for China’s capital market,so as to explore the transmission mechanism of internal capital market and to explore the interaction of characterization vector and the return of capital market.In the course of the study,we find that the characterization vectors and stock returns are not necessarily always showing a linear correlation,and more often,their relationships are complex and nonlinear.To solve these problems,we will in the field of mathematics in the support vector machine theory and stock market data model combined with,support vector machine(SVM)theory is introduced in the quantitative investment stock selection process,established the reaction for the characterization of the vector and the stock return nonlinear hyper plane,makes the relationship between the two measurement more accurate.Based on the established quantitative stock selection model based on support vector machine.The fifth step,we use data from the Chinese stock market,of different markets,different industries,different holding the fitness of quantitative selection model under the term of empirical research for China market in quantitative investment model of investment logic and application depth are discussed.We more than a share market as the research object,for different holding period to quantify the selection model in different industries,different style section adaptability conduct an empirical study.Study,the selection of the data of large amount of samples,mainly above a share market as the research object,through the data screening totaling get 810 stocks in 1666 trading day total 8,096,760 daily data,3,670,531,200 points of the data.Through a large number of empirical analysis,we draw the following conclusions:1.there is a strong correlation between the stock of capital market and stock returns,so it is able to analyze the stock investment portfolio through the analysis of the stock characterization factors.But their relationship is not linear.2.through the introduction of support vector machine theory,the establishment of the relationship between the stock of the reaction stock and the stock of the super plane,you can effectively reflect the relationship between the stock characterization vector and revenue.3.the quantitative analysis model of the quantitative investment representation vector can be used in the quantitative analysis of the different stocks,and the effective screening of the characterization vectors of the reaction stock returns.4.this paper established based on support vector machine theory of quantitative selection model for the Shanghai A-share market as a whole has good adaptability,basically can achieve stock pros and cons of the combination of classification and the equal weight held to realize the objective of income.But according to the different stages of the market and holding period,model adaptation is not consistent,specific showing a,medium-term and long-term holders is superior to that of short-term hold;stock market turbulence and downward market stage effect is better than that of the upstream market.5.different market segments,quantitative stock model showing different effects,in the field of industrial and commercial areas,model of stock classification effect is better,however,model in the field of commercial degree of adaptation to better than the industry;in the field of public utilities and in general,model of stock classification results are relatively poor,formed by a portfolio of shares no significant distinction.6.under the different market style division,quantitative stock selection model for small and mid market stock classification results to outperform the market.But in different market segment,model of downward market and market shocks of the stock classification is better than the upstream market.In general,based on the adaptive analysis of the traditional quantitative models,this paper finds two decisive factors which determine the suitability of the quantitative model.And then put forward the quantitative investment characterization vector effectiveness analysis model and built based on the theory of support vector machine(SVM)quantitative stock selection model on the basis of.Empirical studies have indicated that the model has a good adaptability to the whole A stock market.At the same time,through the introduction of the support vector machine theory in the hyper plane classification method,which provides a useful attempt to study the relationship between stock characterization vector and stock returns from the nonlinear point of view,and has achieved good results.Through quantitative investment model in China’s stock market adaptability analysis,quantitative investment model adaptability evaluation framework is established,and according to the characteristics of the stock market of our country to establish the effectiveness of quantitative investment characterization vector analysis model and the support vector machine(SVM)quantitative selection model was established based on the combined with the nonlinear characteristics of the characterization vector and stock returns.To achieve the theoretical research and empirical research two aspects of innovation,specific as follows:In terms of theory,1.from three angles of value,momentum,technology on the stock market reflect vector for the characterization of the stock returns are sorted and classified,the establishment of quantitative investment characterization vector index system;2.the establishment of quantitative investment characterization vector effectiveness analysis model,make use of the validity of the information ratio to characterize the vector;3.will support vector machine theory and stock selection model combining,established based on support vector machine quantitative stock model,reaction for the characterization of the vector and the stock returns the optimal hyper plane,to stock market stock selection selection and classification.In the empirical research,the establishment of quantitative investment model adaptability evaluation standard,on the basis of comparison of the capital market of our country market is different,different industries,different holding period amount model validity and application of the depth,from multiple dimensions to model in the analysis of the applicability of were studied provides a judgment and evaluation method of the application and development of quantitative investment in the domestic environment.At the same time,the deficiency of this paper lies in the following two points:1.Because the text is based on the commonly used representation vector analysis,the index of the degree of coverage is insufficient.At the same time,because of the sensitivity of the capital market of the state feedback,sensitivity test of the model in this paper by using the inverse effect of the market remains to be studied;2.Due to the complexity of the real world capital market,stock market characterization vector and income between your relationship is often difficult to precisely interpret,based on hyper plane theory of nonlinear relationship does not completely explain it,so to the nonlinear relationship research and measurement needs further,in order to more comprehensively and precisely the relationship between quantitative characterization of stock vector and stock returns,and establish a more effective quantitative investment model. |
PDF Full Text Request