Linear Interpolation and Extrapolation | 线性插值与外推方法及注意事项

📚 Linear Interpolation and Extrapolation | 线性插值与外推方法及注意事项

In IB Mathematics, interpolation and extrapolation are essential techniques for estimating unknown values from known data. They are used in the Applications and Interpretation (AI) course and appear in the Analysis and Approaches (AA) course when studying functions and statistical models. Understanding the methods, their assumptions, and their limitations is critical for scoring well in Paper 2 and Paper 3 questions that involve data modelling and prediction.

在IB数学中，插值与外推是依据已知数据估计未知值的关键技术。这两类方法在“应用与解释”（AI）课程中尤为常用，同时在“分析与方法”（AA）课程研究函数与统计模型时也会出现。理解其计算方法、假设前提以及局限性，对在Paper 2和Paper 3中应对数据建模与预测类问题至关重要。

1. Core Concepts | 核心概念

Interpolation is the process of estimating a value that lies within the range of the given data points. Suppose you have a set of x-values from x₁ to xₙ; any estimate made for an x-value between x₁ and xₙ is an interpolation. Because the estimate stays inside the known data interval, the uncertainty is generally small, provided the underlying trend is consistent.

插值是指在已知数据点范围内估计一个数值的过程。假设你有一组x值，其范围为x₁到xₙ；只要估计的x位于x₁与xₙ之间，就属于插值。由于估计值保持在已知数据区间内部，只要底层趋势一致，不确定性通常较小。

Extrapolation, in contrast, is the process of estimating a value outside the known data range. This happens when you extend the trend beyond the smallest or largest observed x-value. The risk is significantly higher because you assume that the pattern continues without change beyond the data.

外推则是在已知数据范围之外进行估计的过程。当趋势线被延伸到最小或最大观测x值之外时，就产生了外推。其风险显著更高，因为你假设数据范围以外的模式会原样延续。

2. The Data Model Context | 数据模型背景

In IB statistics, interpolation and extrapolation usually involve a regression line of the form

y = ax + b

where a is the slope and b is the intercept. The regression line summarises a linear relationship between two variables, typically obtained using the least squares method. Once you have this equation, you can substitute any x-value to predict the corresponding y-value.

在IB统计学中，插值与外推通常基于形如

y = ax + b

的回归直线，其中a为斜率，b为截距。回归直线通过最小二乘法总结两个变量之间的线性关系。一旦求出该方程，你就可以代入任意x值来预测对应的y值。

The key restriction is that the x-value you substitute must be based on the original domain. If x lies between the minimum and maximum observed x-values, you are interpolating. If x lies outside this interval, you are extrapolating.

关键限制在于：你所代入的x值必须依据原始数据的定义域。如果x位于最小和最大观测x值之间，你就是在插值；如果x位于该区间之外，则是在外推。

3. Linear Interpolation Using Two Points | 利用两点进行线性插值

When you only have two data points, or when you choose two points from a table, linear interpolation assumes the relationship between them is a straight line. The formula for estimating y at a given x between x₁ and x₂ is

y = y₁ + (y₂ − y₁) × (x − x₁) ÷ (x₂ − x₁)

当你只有两个数据点，或者从表格中选择两个点时，线性插值假设它们之间的关系是一条直线。对于介于x₁与x₂之间的x，估计y的公式为

y = y₁ + (y₂ − y₁) × (x − x₁) ÷ (x₂ − x₁)

This formula is essentially the equation of the line joining the two points. It is exact when the data does follow a linear trend, but it becomes an approximation if the true relationship is curved. In IB questions, a linear interpolation often appears when a table gives a function value at two known x coordinates and you must estimate a value in between.

该公式本质上是连接两点的直线方程。当数据确实呈线性趋势时，结果是精确的；但如果真实关系是弯曲的，它就只能是一个近似值。在IB题目中，当表格给出两个已知x坐标的函数值时，线性插值常常用于估计介于两者之间的值。

4. Regression-Based Prediction | 基于回归的预测

The most common IB method is to use the regression line y = ax + b. Your GDC can calculate a and b from a table of data, along with the correlation coefficient r. The regression line minimises the sum of squared vertical distances between the points and the line, making it the best linear fit in a statistical sense.

最常见的IB方法是使用回归直线y = ax + b。你的图形计算器（GDC）可以从数据表计算a和b，同时给出相关系数r。回归直线使数据点与直线之间的垂直距离平方和最小，因此它是统计学意义上最佳的线性拟合。

To interpolate, choose an x-value inside the observed range. For example, if x ranges from 0 to 100, and you predict y at x = 45, that is an interpolation. The regression line provides a reasonable estimate, assuming the linear trend holds convincingly across the whole interval.

要进行插值，只需在观测范围内选择一个x值。例如，若x范围是0到100，而你预测x = 45处的y值，这就属于插值。只要线性趋势在整个区间内稳定成立，回归直线就能给出合理的估计。

To extrapolate, choose an x-value beyond the maximum or below the minimum. For instance, using the same data, predicting y at x = 150 would be an extrapolation. The prediction rests on the assumption that the linear relationship remains valid outside the data range, which is rarely guaranteed.

要进行外推，则需选择超出最大值或低于最小值的x值。例如，使用同一组数据，预测x = 150处的y值就属于外推。该预测依赖于“线性关系在数据范围之外仍然成立”的假设，而这种假设很少能保证。

5. Non-Linear Models and Curve Fitting | 非线性模型与曲线拟合

Not all IB data are linear. For exponential growth or decay, you might fit a model of the form

y = abˣ

where b is the growth or decay factor. For quadratic relationships, you might use

y = ax² + bx + c

并非所有IB数据都是线性的。对于指数增长或衰减，你可以拟合形如

y = abˣ

的模型，其中b是增长或衰减因子。对于二次关系，则可以使用

y = ax² + bx + c

When you use these models, the same interpolation/extrapolation logic applies. Interpolating inside the data range is usually safe if the model fits well; extrapolating beyond the range is dangerous because exponential curves can quickly produce unrealistic predictions. For example, a population model might predict reasonable values for the next few years, but extrapolating 100 years into the future could produce absurdly large numbers.

使用这些模型时，同样的插值/外推逻辑依然适用。只要模型拟合良好，在数据范围内插值通常是安全的；但超出范围外推则风险很高，因为指数曲线可能迅速产生不现实的预测。例如，一个人口模型也许能合理预测未来几年的数值，但外推100年后可能会得到荒谬的巨大数字。

6. Comparison | 插值与外推对比

Aspect	Interpolation	Extrapolation
x-value position	Inside observed interval	Outside observed interval
Reliability	Higher, if trend is consistent	Lower, assumes continued pattern
Main assumption	Model holds across the gap	Same trend extends infinitely
Typical IB use	Estimating missing table values	Forecasting future values

The table above summarises the key differences. In IB markschemes, examiners award credit for correctly identifying whether an estimate is an interpolation or an extrapolation, because it affects the reliability of the answer.

上表总结了关键区别。在IB评分方案中，考官会对“正确判断估计是插值还是外推”给予分数，因为这直接影响答案的可靠性。

7. Dangers and Limitations | 风险与局限性

The biggest danger in extrapolation is the assumption of the continuation of the trend. Real-world data often follow a pattern that changes after a certain point. For example, a linear regression of temperature against time might look good over ten years, but extrapolating fifty years in the future could predict impossible temperatures. The correlation coefficient r only measures the strength of the linear relationship within the observed data; it says nothing about outside data.

外推的最大风险在于假设趋势会继续延续。现实世界的数据往往在某个节点后发生变化。例如，温度-时间线性回归也许在十年内看起来很好，但外推五十年后可能会得到不可能的极端温度。相关系数r只衡量观测数据内的线性强度，不能说明数据范围外的情况。

Another limitation is that interpolated values are not guaranteed to be correct either, especially if the gap between two data points is large. If the true relationship is curved but you use linear interpolation, your result will be biased. Always check for a visual trend in a scatter plot before applying a method.

另一个局限是，插值结果也不能保证完全正确，特别是当两个数据点之间的间隔较大时。如果真实关系是弯曲的，而你使用线性插值，结果会存在偏差。在应用任何方法之前，务必先观察散点图的整体趋势。

Furthermore, extrapolation can produce values that are practically impossible, such as negative population sizes or negative prices. In IB exams, you are expected to mention this when discussing the validity of predictions.

此外，外推可能产生实际不可能的值，例如负的人口数量或负的价格。在IB考试中，讨论预测的有效性时，你应当主动提到这一点。

8. IB Exam Context and GDC Skills | IB考试情境与计算器技巧

In IB Mathematics AI Paper 2, you will often be given a table of data and asked to perform a linear regression. The command term may be “estimate”, “predict”, or “comment on the reliability”. You should be comfortable with the following GDC steps:

在IB数学AI Paper 2中，你经常会得到一组数据表，并被要求进行线性回归。指令词可能是“估计”“预测”或“评价可靠性”。你应该熟练掌握以下GDC步骤：

Enter the data into two lists (e.g., List 1 and List 2).

将数据输入两个列表（例如列表1和列表2）。
Calculate the least squares regression line using the statistics menu.

使用统计菜单计算最小二乘回归直线。
Record the values of a (slope), b (intercept), and r (correlation coefficient).

记录a（斜率）、b（截距）和r（相关系数）的数值。
Substitute the required x-value into your equation, or use the graph and trace functions.

将所需x值代入你的方程，或使用图形和追踪功能。

For non-linear models, use the exponential or quadratic regression options. Always check whether the resulting model passes through the general cluster of data points and whether the residual pattern looks random.

对于非线性模型，使用指数或二次回归选项。始终检查所得模型是否穿过数据点的整体簇团，以及残差模式是否呈现随机性。

9. Common Pitfalls | 常见错误

Students frequently confuse interpolation with extrapolation. To avoid this mistake, explicitly state the observed minimum and maximum x-values in your working. If your prediction x-value lies between those two bounds, say “this is interpolation”; otherwise, say “this is extrapolation”.

学生经常混淆插值与外推。为避免这一错误，请在解题过程中明确写出观测x值的最小值和最大值。如果你的预测x值位于这两个边界之间，就说“这属于插值”；否则就说“这属于外推”。

Another common pitfall is using a regression line without verifying whether the linear model is appropriate. If the scatter plot suggests a curve, linear interpolation will give meaningless results. In the exam, you may be asked to state an assumption or limitation of the model, and you must mention the danger of extrapolating beyond the given data range.

另一个常见错误是不验证线性模型是否合适就直接使用回归直线。如果散点图显示弯曲趋势，线性插值将产生无意义的结果。考试中，你可能会被要求说明模型的一个假设或局限，此时必须提及超出给定数据范围进行外推的风险。

Finally, do not over-round intermediate values. Keep at least three significant figures for a and b during calculation, and round only the final answer to the required degree of accuracy.

最后，不要过早四舍五入中间值。在计算过程中，a和b至少保留三位有效数字，只对最终答案按题目要求进行舍入。

10. Worked Example | 完整例题

A farmer records the height of a plant (h cm) at different times (t weeks):

一位农民记录植物高度h（厘米）随时间t（周）的变化数据：

t (weeks)	1	2	3	4	5
h (cm)	8	12	18	25	33

Using the least squares regression line, estimate the height when t = 3.5 and when t = 7. State which estimate is interpolation and which is extrapolation.

使用最小二乘回归直线，估计t = 3.5与t = 7时的植物高度，并说明哪个是插值、哪个是外推。

The regression line calculated from the data is

h = 1.2t⁺6.4 (approximately)

由数据计算得到的回归直线为

h = 1.2t + 6.4（近似）

For t = 3.5:

h = 1.2(3.5) + 6.4 = 4.2 + 6.4 = 10.6 cm

Since 3.5 lies between 1 and 5, this is interpolation.

对于t = 3.5：

h = 1.2(3.5) + 6.4 = 4.2 + 6.4 = 10.6 cm

因为3.5介于1和5之间，所以这是插值。

For t = 7:

h = 1.2(7) + 6.4 = 8.4 + 6.4 = 14.8 cm

Since 7 is greater than 5, this is extrapolation, and the prediction is unreliable because we have no evidence that the linear trend continues beyond t = 5.

对于t = 7：

h = 1.2(7) + 6.4 = 8.4 + 6.4 = 14.8 cm

因为7大于5，所以这是外推，且该预测不可靠，因为我们没有证据表明线性趋势在t = 5之后仍然延续。

11. Summary and Revision Tips | 总结与复习建议

Interpolation and extrapolation are powerful tools in IB Mathematics, but they are only as valid as the model underlying them. Interpolation is safer because it stays within the observed data range; extrapolation requires strong justification and is often the subject of exam questions that test your ability to evaluate assumptions.

插值与外推是IB数学中的强大工具，但它们的有效性完全依赖于底层模型的合理性。插值因保持在观测数据范围内而更安全；外推则需要强有力的理由支撑，并且常被用作考试题目来考查你评估假设的能力。

Always write down the regression equation with at least three significant figures.

始终写出至少保留三位有效数字的回归方程。
Clearly state whether a prediction is interpolation or extrapolation.

清楚说明某个预测是插值还是外推。
Mention at least one limitation when asked to comment on reliability.

当被要求评价可靠性时，至少提到一个局限性。
Check that the correlation coefficient r is strong enough to justify a linear model.

检查相关系数r是否足够强，以支持线性模型。
Use a scatter plot on your GDC to visually confirm the trend before applying regression.

在应用回归前，用GDC绘制散点图直观确认趋势。

In revision, practise past paper questions that involve tables and predictions. Focus on writing short justifications, such as “this is an extrapolation, so it is not reliable because the trend may not continue outside the data range.” This precise wording will earn you method marks.

复习时，多练习涉及数据表和预测的历年真题。重点练习写出简短理由，例如“这是外推，因此不可靠，因为数据范围之外的趋势可能不会延续。”这种精确表述可以帮你获得过程分。

Published by TutorHao | IB Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Linear Interpolation and Extrapolation | 线性插值与外推方法及注意事项