129770904179843750_169[Multi-factor model is the most important category of quantitative stock selection model, the basic idea is to find and return the most relevant indicators. And according to the index, build a stock portfolio, expect the combination of beat for some time in the future or running distance index. If the beat, you can do more of the combinations and short period, earning positive alpha return; if you are runningTransmission, you can go long periods, margin shorting the portfolio, make reverse Alpha returns. Multi-factor model is the key to find the association between factor and yields] basic concepts of multi-factor model is the most widely used a stock selection model, the basic principle is the use of a range of factors as stock selection criteria, these factors are met shares were bought, sold does not meet theOut. Take a simple example: If a group of people to participate in a marathon, want to know what people are going to come to above average, it just before you run a physical test can be done. Health indicators by the former athlete, average grades were more likely to get beyond. Similar principles of multi-factor model, we only have to find factors that yield the most relevant to the enterpriseCan be.����Multi-factor model of core differences on the selection of the first factor, the second is how to be a multiple factor comprehensive final judgment.����In General, the method of multiple-factor selection model has two kinds of judgment, is a scoring method, the second is a regression. Marking method is based on the size of each factor to rate the stock, and then weighted according to certain weightTo a total, according to the total score and then filter the stock.����Yields on stocks of the regression method was used in the past to return to a number of factors, get a regression equation, and then factor the latest value in the regression equation are a stock returns predict the future, and then based on this selection. Multiple-factor selection model building process can be divided into the factor of candidates selectedTaking, testing the effectiveness of stock selection factor, effective removal of, comprehensive redundancy factor score model and model evaluation and continuous improvement of the 5 steps. Factor candidate of choice depends largely on the economic logic of the selected candidate factors and market experience, but more and more effective factor is undoubtedly enhanced model of information capture capabilities, increase profitability of one of the key factors�� For example: on January 1, 2011, select circulation market value maximum of 50 stocks, building a portfolio, held to the end of 2011, the combination can achieve yields in excess of 10%.����It just goes to show that in 2011, this time, circulation market value and ultimately there is a positive correlation between the rate of return. From this example you can see that the most simpleMulti-factor model describes a relationship between factor and yield on some time in the future. Similarly, you can select a different factor, for example, may be some fundamental indicators, such as PB, PE, EPS growth rate and so on, may also be a number of technical indicators, such as momentum, such as fluctuations in exchange rates, or other indicators, such as earnings growth, analyst consensus estimateOf macroeconomic variables
diablo 3 gold, and so on.����Same time period, is also an important parameter, was held for a month, two months, or a year, to the end yields great. Stock selection factors test the effectiveness of inspection method mainly uses the sort method of the test candidates in general factor of stock selection effectiveness. For example: monthly examination, specific andStatements, for any a a candidate factor, in model formed period of first a month began calculation market in the each only normal trading stock of the factor of size, by small to large of order on sample stock for sort, and average is divided into n a combination, has been holds to month, next month then by also of method again construction n a combination and holds to month, monthly so, has been repeat toModels formed the final. Examples above have explained this test method, also can test for the nth month, 2 months, 3 months
swtor credits, or even longer. There is also a parameter is the number of combinations of candidates, are 50, or 100, are very important parameters.����Optimal selection of specific parameters, you need to use historical data to test. There areBut redundancy removal of different factors of stock selection factor that may be due to the inherent driver of much the same reasons, elected by the combination in terms of stocks, and income have a higher consistency, so some of the factors which need to be removed as redundant factor, while keeping the same factor income in one of the best, the most distinguish factor. For example, indicators of turnover and liquidityObvious correlation between indices.����Greater circulation of volumes are generally large, so the stock selection model, these two factors to select only one of them. Removed redundancy factor method: suppose you want to select effective factor k, sample period m months, then removed redundancy factor of concrete steps to: (1) under different factors on n GroupEvaluation, score and the combined throughout the model form of income-related, income greater, score higher, (2) monthly correlation matrix between the stocks of different factor score, and (3) after you calculate the monthly correlation matrix factor score, calculate the average correlation matrix in the entire sample period, and (4) set a threshold score correlation between mInScoreCorr, correlation between scoring average is greater than the threshold corresponding to the elements in the matrix factor only and other small factors, efficiency of more factors, and other factors are removed as redundant factor. Establishment of comprehensive evaluation model and select after removal of redundant stock selection of comprehensive evaluation model of effective factors in the model runs aTime, for example, each month, for normal transactions in the market share for each factor of the latest scores and evaluated in accordance with a certain amount of weight the average score of all factors. Finally, depending on the model of the comprehensive sort the stock average reached at, and then select top stocks as required. For example, select the 20% before the stock with the highest score, or select the highest scores50 to 100 stocks. For example: you can build a one-factor model is (PE,PB,ROE), when earlier this month, several factors on the rate, then 50 stocks with the highest score as a portfolio, take turns replacing the next month by the same method. Persist for some time, examination of the portfolio yieldsBeat the baseline, which is the establishment of integrated scoring model and a posteriori.����Of course, this example is one of the most simple example, actual model may be more complex, such as Volvo score is a complex multi-factor model, it is to trade the stock, a combination of high scores for each industry
diablo 3 power leveling, and then into the investment baskets. ModelEvaluation and continuous improvement on the one hand, due to the volume of stock selection method is invalid or weak premise established in the market, as the number of investors who use multiple-factor selection model continues to increase, some factors will eventually fail, and some new factors may be verified effectively and to model; on the other hand, some factors might be in the market environment of the pastEfficiency, and with the style change in the market, and they may fail in the short term, and some previously invalid factor in the current market environment better. In addition, calculations in the grading process, the weight of each factor score design, taking into account transaction costs and risk control, there is room for further improvement. Preferential stock selection model in the use ofFactor, the selection of the model itself continued evaluation and continuous improvement to adapt to the changing market environment. Model two is the most important aspect of multi-factor: a is valid factor, is another factor of the parameter. For example, PE is valid or ROE effective; 1 month exactly is do storage cycle was 3 months and bin regulation period. These factors and parametersGet can only be obtained through historical back-test. But back in the measuring process, note that the optimization cannot be excessive, otherwise the results may be bad.
Others:
No comments:
Post a Comment