Use "dataset" in a sentence
dataset example sentences
dataset
1. A representation in CMMI is analogous to a view into a dataset provided by a database
2. One of the best ways to do this is to “feed” an indicator an artificially created dataset that focuses on specific types of market action
3. When thousands of academics and practitioners study the same historical dataset, all motivated to find profitable strategies, they (we) are bound to uncover many empirical relations that just happened to work well in sample and fail miserably out of sample
4. • Besides out-of-sample analysis, other ways to mitigate overfitting bias include requiring a certain economic logic from regularities (incorporating economic priors is useful but in reality they too reflect our experience); conducting cross-validation exercises (testing findings on a dataset other than the one used for fitting the model); extensive robustness checks (subperiod consistency, etc
5. (I use yet another dataset in my 2003 Journal of Portfolio Management article that discusses these issues in much more detail
6. Dynamic strategies give us many more degrees of freedom because numerous small choices (each of which may marginally improve success in a backtest) are made by one researcher repeatedly examining the same dataset, or by different generations of researchers doing the same [1]
7. Even if some performance persistence across specifications exists, there is an obvious danger of overcrowding when many trend-followers make similar decisions using the same dataset
8. Research we have conducted using the MSCI EAFE dataset (Europe, Australia and Far East) shows the strategies performing with a level of excess returns similar to those in the United States
9. For this edition of this book, we use two datasets—the Standard & Poor’s Compustat Active and Research Database from 1963 through 2009 and the Center for Research in Security Price (CRSP) dataset from 1926 through 2009
10. The CRSP dataset provides U
11. Both Compustat’s and CRSP’s research files include stocks that were originally listed in the dataset but were removed because of merger, bankruptcy, or some other reason
12. For example, at the end of 2009, of the 6,705 stocks in our dataset, more than 2,555 stocks were jettisoned because their market capitalization fell below an inflation-adjusted minimum of $200 million
13. Thus, while it is easy to assume that you could purchase and sell these securities at their listed price in the historical dataset, I believe that this is an illusion and unnecessarily gives an upward bias to the results of studies that allow their inclusion
14. One potential problem with the earlier data is the changing nature of the Compustat dataset
15. In this new edition of this book, we use the CRSP dataset to look at stocks by market capitalization between 1926 and 2009
16. Large Stocks are those with a market capitalization greater than the dataset average (usually the top 17 percent of the entire dataset by market capitalization)
17. Finally, we use the Compustat dataset to look at a universe of large capitalization stocks composed of market-leading companies
18. We need the Compustat dataset for our market-leading companies—which I call Market Leaders—because several of the factors we use to create Market Leaders are not available in the CRSP dataset
19. Included with the Compustat data are the data from the CRSP dataset, leading me to conclude that if you could buy these tiny names, the most realistic average annual returns would be between 17
20. When you extend the analysis to include the period from July 1926 through December 2009 using the CRSP dataset, the average annual compound return for noninvestable microcaps falls to 15 percent per year
21. Using the Compustat dataset, between 1964 and 2009 investable microcap stocks earned an average annual compound return of 12
22. 70 percent, whereas when using the CRSP dataset over the same period microcap stocks earned 11
23. When you use the CRSP dataset to review the full period between 1926 and 2009, the investable microcap stocks compound at 10
24. But that was only 18 years of data, and as we have seen, such a small dataset limits your ability to draw true conclusions as to how robust the results actually are
25. While I do not have payout data in the CRSP dataset, when studying stocks with high-dividend yields, I do have the data in the Compustat dataset, starting in 1963
26. 5 times the dataset average (50 percent higher than average)
27. Since we are able to use the CRSP dataset to calculate buyback yields, we begin on December 31, 1926, and invest $10,000 in the deciles of stocks with both the highest and lowest buyback yields from both the All Stock and Large Stocks universes
28. Since we have the data we need to conduct this analysis in the CRSP dataset, we start with a $10,000 investment on December 31, 1926, and hold it through to December 31, 2009
29. This remains largely true even when you use the CRSP dataset to analyze an additional 37 years of data
30. While we don’t have volume statistics available in the CRSP dataset, we do have them in the Compustat dataset beginning in 1964
31. 25 summarize the returns and base rates for a variety of multifactor strategies using the CRSP dataset
32. This combination of value and growth characteristics is one we revisit when we look at even more multifactor models using research from the Compustat dataset
33. Applying these factors to the Compustat dataset qualifies just 6 percent of the stocks as Market Leaders
34. 2 compares the universes’ performance for the Compustat dataset, which covers the 1963 to 2009 time period
35. For example, when examining some of the strategies with the longer-term CRSP dataset, we find several factors that do not earn those 100 percent scores, that is, the five-year base rate for Small Stocks with the best buyback yield drops to 89 percent, and the ten-year base rate to 87 percent
36. For two of the best performing strategies—buying Small Stocks with the best buy-back yield and buying Small Stocks with the best six-month price appreciation—we are able to further verify their outperformance by looking at the CRSP dataset between 1926 and 1963, where the Compustat dataset begins, thus adding 37 years of confirming data
37. This is also the result of the additional data we were able to study from CRSP dataset
38. And because we now have access to the CRSP dataset, we’ve also discovered that six-month price appreciation is a more effective final momentum filter than 12-month price appreciation
39. You will recall that for a stock to be included in the Market Leaders universe, it must be a nonutility company that has a market capitalization greater than the average of the Compustat dataset; cash flow greater than the average; shares outstanding greater than average; and annual sales 50 percent greater than the average for the dataset
40. We have run tests similar to those in this book on the MSCI dataset that begins in 1970 and found that, for the most part, these strategies work equally well in foreign markets
41. This, by itself, would encourage us to focus on an early period of our dataset, perhaps the earliest year of 1995
