[Network Report] The mechanism of selecting stock logic gene variation quantitative investment in internet excavation king data

21st Century Shenzhen, Ningxia report

Using Internet financial big data to build stock-selecting model is becoming a new trend in a-share market.

October 20, the GF Fund and Baidu co-development of the GF in the certificate Baidu Hundred Strategy 100 Index fund will start the release, this is the industry's first track with the Internet gene Index of the index-type fund products.

Plan to launch Internet gene

Related Company stock trend

• China Merchants Securities 32.95+1.163.65%
• Changjiang Securities 16.52+0.493.06%

Index products are also the Southern Fund Management Co., Ltd.

"The Fund product programme, which tracks the Big Data Index, will be formally declared by regulators in the near future and is expected to issue fund products that track the i100 index during the year." "South Fund product Development Department deputy director Li revealed.

According to the 21st century economic reporter interviewed that, in addition to GF, South Fund, the industry also has a fund company's quantitative investment department and dedicated to research "big data" quantitative strategy.

In addition, the financial engineering team at the Brokerage institute is also tapping into "big data" investment opportunities. The industry representatives of the Yangtze River Securities financial engineering team, they early in the year, according to Sina Financial news data to build a news stock selection model.

The glamour of Big data

Using Internet big data to excavate the stock market's excess income opportunity is becoming a new trend in the recent fund industry. And the action is relatively fast as the GF Fund and the South Fund of the two fund companies.

whichGF Fund Joint Baidu Company, the Chinese Certificate Index company Development Hundred 100 Index, the southern Fund and Sina Finance, Shen Certificate information company launched the i100 Index and i300 index.

The incentive for fund companies to compete in the Big Data index is to outperform existing index funds, based on a combination of big data filters.

According to the data provided by the index company, the cumulative yield of the Hundred 100 Index reached 545% from 2009 to 2014 June 30. In the same period, the yield of the 500, medium and CSI 300 indices was 102%, 56% and 19% respectively.

Historical gains are derived from simulations of model samples, and there may be doubts about their profitability. And the Hundred-Fat 100 index product in the model sample outside of the real data, the same significant run to win the mainstream index.

GF Fund Quantitative investment researcher Jifeng to reporters, beginning June 20, hundred 100 index into the real plate operation stage. From this to October 8, the Hundred-Fat 100 index achieved a cumulative yield of 43.33%. In the same period, the Shanghai and Shenzhen 300 Index, Shanghai index and Gem are only 16.52%, 17.74% and 17.94% of the yield respectively.

The historical yield of weight indices such as i100 and i300 also highlights the advantages of big data.

From January 29, 2010 to July 31, 2014, the cumulative yield of i100 and i300 reached 222.4% and 141.58% respectively, well above the cumulative earnings of the gem index of 34.45%, more than the cumulative yield of the SME Board Index-8.95%.

Liu Jiping, director of the Southern Fund's Quantitative investment division, introduced the southern Sina Big Data Index, which began running outside the model sample since August. Among them, the i100 index in August and September to achieve the monthly yield of 10.93%, 15.63% respectively. This means that the index has a yield of 26.56% over a two-month period.

The brokerage's research team is also tapping into the investment opportunities that "big data" brings. whichThe Changjiang Securities Financial Engineering Team established the news stock selection model since the beginning of the year, sample tracking is starting from January 12.

The research report published by Fan Xinting, head of the Changjiang Securities Finance project, showed that the total yield of the news stock selection model was 52.45%, exceeding the CSI 300 index of 39.44% and exceeding the 500 index 30% in January 22 to August 15.

　　 Xia Yangyang, senior analyst of China Merchants Securities Financial Engineering, used the interactive and easy-to-disclose research information of SZSE to build SME Board GEM Research portfolio. The results of real disk tracking show that from the beginning of the year to September 30, the cumulative yield of the research portfolio was 29.31%, the SME Board index was 14.91%, and the margin beyond the gem index was 8.53%.

Introducing the Internet gene

Whether it is the stock selection model of Changjiang Securities and China Merchants Securities, or the Big Data Index launched by GF Fund and South Fund, its common pointsis to introduce netizens to search big data of a stock as the stock-selecting factor.

The principle of the hundred-Fat 100 index is to increase the search factor of Baidu Big data on the basis of the traditional index model, which is equivalent to the early behavior of investors into index indicators. "Simon Luk, general manager of the Investment Department of GF Fund, said.

Hundreds of 100 of the participation of developers, GF Fund quantitative Investment Research researcher Jifeng, hundred 100 model is also called the investor comprehensive sentiment model, it is equivalent to a large data furnace, the financial factors, momentum factor, search factors and other types of data to do a long dimension analysis, so that it in the high-dimensional space chemical reaction.

"Using Baidu platform data, the stock of the sample space for the last one months of the total search and search increment, respectively, counted as the total factor and increment factor." "Jifeng Introduction, Baidu Financial big data, in a sense is a measure of investor sentiment, one of the quantitative methods, can be used to reflect the investor's investment will or expectations."

"The innovation of the Hundred-Fat 100 index is that the network search volume has become the factor of choice components, the index of the construction of the use of Baidu in the search field of strong resources, with significant internet genes." "Liao, a researcher at the research Center of Zhong Lu Fund.

The information disclosed by the company's website shows that the sampling method of the hundred-Fat 100 index is arranged in descending order according to the comprehensive score of the comprehensive financial factor, the comprehensive momentum factor and the search factor, and selects the top 100 stocks as the constituent stocks of the Hundred 100 Index.

Among them, the comprehensive financial factors include 8 financial factors, such as the return on net assets, the rate of return and the growth of earnings per share. The integrated momentum factor is the calculation of the stock price yield and volatility in the last one months, and the momentum index after the risk adjustment is obtained. According to its size from high to low, ranked in the first 30% of the momentum factor, the latter 30% is recorded as a reversal factor.

The southern Sina financial Big Data sampling method also introduced the domestic first door home Sina network Big data.

The "I" Index's stock-selecting factors include financial factors, big data factors, and market-driven factors. Field Museum, director of the Southern Fund's Equity and investment division, said the big data factor covers Sina Finance'spage hits, positive and negative tweets, news reports.

In addition, financial factors include the latest P/E ratio PE, ROE, annual operating income and net profit growth rate, and their change forecasts. Market driving factor refers to the stock turnover rate, volatility, price change rate, volume change rate.

Long before the South Fund, the financial engineering team of Changjiang Securities used Sina Financial data to construct the news stock selection model.

"We selected the top 50 stocks of the fastest-growing weekly click-through to Sina's financial data, excluding companies with bad news and building a portfolio of 50 stocks. Fan Xinting points out that the core of the news stock selection model is to select a stock construction combination which has a great increase in recent clicks, in order to find the expected changes brought about by the news heat.

Although the introduction of Internet financial big data as a stock selection factor, but different institutions to build the constituent stocks still have obvious difference.

According to the information disclosed by the China Securities Index company, the 100 Index of the first 100 constituent stocks, the sample shares in each market segment relatively dispersed. Among them, SME board shares the largest number, reached 39. The second is the Shanghai Main Board 31, the deep License board and GEM each accounted for 25 and 5.

From the industry distribution, the Hundred 100 Index of the constituent stocks in the top five industries are industry, optional consumption, raw materials, financial real estate, information technology, the proportion of the index of the weight ratio reached 85%. Compared with the CSI 300 index, its financial real estate allocation ratio is relatively small, information technology and optional consumption and so on obviously over.

According to financial factors, big data factor and market driving factor, the constituents of the I index are mainly distributed in four industries, such as raw materials, information technology, industry and optional consumption. Among them, the i100 index, the i300 index in the above four industries weight accounted for 73% and 71% respectively.

Subversion of stock selection logic

When the fund company and the securities firm study the big data of internet finance as the stock selection factor, it represents the asset management organization to reconstruct the stock selection logic in the index investment.

"Traditional indexing relies on traditional factors such as market capitalization, transaction amount, financial and valuation." Its biggest disadvantage is to use the last 3 months or 6 months of data to predict the future period of revenue, parameters of historical data have a serious dependence. Simon Luk pointed out that the use of Baidu data into the investors in the investment decision before the behavior of the law, the future market investment law has a certain predictive effect, its prediction is better than the traditional historical data from the factor data.

In contrast to the traditional index, the idea of the Hundred-100 index is to jump out of the industry, the constraints of the plate, from the whole market to look for excess revenue opportunities.

"We want to try to find the hot spots in the market for the next one months through financial data, momentum data and big data from Baidu. Jifeng that the hundred-Fat 100 index stock model is characterized by its selection of stocks is to match the future market or industry rotation hot spots, and fundamental good, the future has a certain growth space value-based stocks.

The Big Data Index, launched by the Southern Fund and Sina, also uses the same approach.

"Take advantage of big data investment to the market breadth to gain excess." "From 2005 to 2014, even if the average yield in the market is negative, you can still find 30% of the stock that goes beyond the market," Field Museum, director of the Southern Fund's Equity and investment division, according to wind statistics.

For example, from the beginning of 2014 to August 22, wind full A has a yield of 14.12%, exceeding the market 30% of the number of shares to 535, the proportion of the total market share is 22%. Among them, the average yield of the stock reached 74.63%.

In the industry's opinion, the fund's Quantitative investment department has introduced big data on the Internet, with the aim of digging out the underlying market-average earnings. Based on this, compared with the traditional index half-yearly sample unit, the hundred-Fat 100 index and the i100, i300 index of the sample stock of the frequency of the position is set to every month.

"About 50% of the constituent stocks will be adjusted every month. Jifeng, the Hundred Fat 100 Index in the third week of the month in Friday to carry out a sample warehouse. Since June 20, the sample has been adjusted three times. Of these, about half of the constituent stocks are replaced each month.

Jifeng admits that, compared to other indices, the percentage of stocks in the Hundred-100 index is relatively high. But he argues that the index sample is adjusted more frequently and in a far more traditional way than the conventional index, toto better grasp the market's wheel movement quotes。

Shenzhen Securities Information Co., Ltd. Shiping the data also shows that from July 1, 2013 to June 1, 2014, the southern Sina big data in nearly 12 sample stocks, the i100 index on average 41 stocks per adjustment, the i300 Index average adjustment of 95 samples per unit, The monthly average position ratio is 43.5% and 34.5%, respectively.

In the industry's view, for nearly two years,the domestic market has the characteristic of obvious theme wheel movement. The strategy index developed based on big data is well suited to the market for theme rotations.

However, some research institutions have suggested that the introduction of big data models can withstand the test, still need time to verify.

"The construction of the hundred-Fat 100 index and I coefficient index, while considering the value of a stock, also actively grasp the market focus and public opinion, such a starting point for good performance is beneficial." Liao that using the Internet to build quantitative models of genes is reasonable enough to give the answer in time.

Ningxia

Transferred from: http://business.sohu.com/20141010/n404965780.shtml

[Network Report] The mechanism of selecting stock logic gene variation quantitative investment in internet excavation king data