《数据处理英文版》由会员分享，可在线阅读，更多相关《数据处理英文版（22页珍藏版）》请在金锄头文库上搜索。
1、实验设计与数据处理 Experimental Design and Data Processing,组长：白宇超 组员：董妍芸 柳耀博 庞士磊 邓丽君 张健榕 吴 波,Data Processing,数据处理分析是指用适当的统计方法对收集来的大量第一手资料和第二手资料进行分析，以求最大化地开发数据资料的功能，发挥数据的作用。是为了提取有用信息和形成结论而对数据加以详细研究和概括总结的过程。数据也称观测值，是实验、测量、观察、调查等的结果，常以数量的形式给出。,Aim,Data processing analysis refers to the use of appropriate statis
2、tical methods to collect a large number of first-hand information and second-hand information for analysis, in order to maximize the development of data functions, data play a role. Is to extract useful information and form a conclusion to the data to conduct a detailed study and summary of the proc
3、ess. Data, also known as observations, is the experiment, measurement, observation, investigation and other results, often given in the form of quantity.,Data processing methods,Many methods of data processing: parameter estimation, hypothesis testing, analysis of variance and regression analysis. P
4、arameter estimation is to estimate some data points and interval estimates; Hypothesis test is to determine the reliability of the results of various data processing; analysis of variance is the impact of various factors on the impact of indicators of the degree of significance; regression analysis
5、is how to get things reflect Mathematical expression of objective law.,Number of significant digits,The more digits after the decimal point, the more accurate the data? (How to determine the effective number of data?) Correct approach: the number of bits taken in addition to the last one figure for
6、the measurement of the suspicious number or estimate, the rest of the figures are accurate and reliable. Significant digits: All numbers in a measurement that reflect the size of the measured value with a questionable number (estimated reading). 。,Rounding rule,A set of values to retain n-bit effect
7、ive, in accordance with the “round-off“ method on the n +1 bit processing, will have rounding error. In general, the first n + 1 digits for the 0,1,2,3,4,5,6,7,8,9 probability is equal, if you use the capital letter X said rounding error, various rounding The odds of error are: P（X=0,-1,-2,-3,-4,5,4
8、,3,2,1）1/10 In a large number of data rounding processes, the errors produced by the summing can cancel each other out 。,“Rounding“ method, when the first n +1 bit is 5, the rounding error is 5, this error can not be offset. This is the classical “rounded“ method of abuse. (How to solve?) Chinas Sci
9、ence and Technology Commission formally promulgated the “Mathematics Revise Rules“, known as the “rounding five single and double“ rule: Rounding off into five to consider, five non-zero after one into the one. Five are zero after the parity, even before the five should be left, five before the odd
10、into the one.,Rounding rule,“Rounding five single and double” rule instructions (四舍六入五单双）,1. f the first digit to be discarded is less than 5, the previous digit remains unchanged. 2. f the discarded first digit is greater than 5, the previous digit is incremented by one 。 3. If the first digit to b
11、e discarded is equal to 5, and all subsequent digits are zero, then the last digit to be retained is odd or even, depending on whether it is an odd number or an even number. Into an odd number, even when the homes.,4. If the first digit to be discarded is equal to 5, and not all of the following num
12、bers are zero, then the next. 5. If the number to be discarded consists of several digits, the number shall not be consecutively carried or discarded 。 6. Rounding the integer should also follow the above rules. For example, 23438, only take three significant figures, it should be 23400 or 2.34 104.
13、,“Rounding five single and double” rule instructions (四舍六入五单双）,Determination of significant figures,Accuracy of measuring instruments: consistent with the accuracy of measuring instruments; Effective number of calculations: follow the “advanced homes, after the operation“ principle.,Data Representat
14、ion,Basic requirements: The representation of experimental and production data requires accurate, concise, and image. The main way of data representation: list method, mapping method and empirical formula method.,List method,1, simple and accurate header (name), placed on the table above. The sequen
15、ce number is placed in front of the meter head. 2, a reasonable choice of the items listed in the table. 3, the project should include the name and unit, as far as possible using symbols.,List method,4, the main item in the table represents the independent variable, the deputy item represents the de
16、pendent variable. 5, the number should be uniform and uniform. The same vertical line of figures, the decimal point to the upper and lower alignment. When the number is zero, the number of significant digits is guaranteed. For example, the effective number of digits for the two decimal places, the zero should be counted as 0.00. 6, note that the data source and the table can not be reflected in the need to explain other issues.,表10 酵母对发酵茶汤风味及生物量的影响（ ）,注：用Duncan法进行多重比较，同行标有不同大写字母者表示组间差异极
