Scribd
Upload
32 views15 pages

22-Intro To Inference For Decision Making-19-03-2024

This document discusses key concepts in hypothesis testing and estimation including point and interval estimates, unbiased and efficient estimators, confidence intervals, types of hypotheses and errors, and critical regions and values. It provides examples and explanations of these statistical inference topics.

Uploaded by

nchiva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
32 views15 pages

22-Intro To Inference For Decision Making-19-03-2024

This document discusses key concepts in hypothesis testing and estimation including point and interval estimates, unbiased and efficient estimators, confidence intervals, types of hypotheses and errors, and critical regions and values. It provides examples and explanations of these statistical inference topics.

Uploaded by

nchiva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

Module-4

Inference for Decision Making - I


෠ is used to estimate the value of parameter 𝜃 then 𝜃෠ is called estimator.
A statistic (𝜃)
Point estimate: If an estimate of the population parameter 𝜃 is given by a single value then it is
called point estimate.
Interval estimate: If an estimate of the population parameter 𝜃 is given by two different values
between which 𝜃 is expected to lie, then it is called interval estimation.
i) Unbiased estimator: An estimator 𝜃෠ of the parameter 𝜃 is said to be unbiased if the mean of
sampling distribution of 𝜃෠ is 𝜃. i.e., 𝐸 𝜃෠ = 𝜃.
Ex:
1) The sample mean 𝑥ҧ is an unbiased estimator of population mean 𝜇.
2) The sample variance 𝑆 2 is an unbiased estimator of population variance 𝜎 2 .
ii) Efficient estimator: If there are more than one unbiased estimator for the population
parameter then the estimator with smaller variance is called efficient estimator.
iii) Maximum error:
For large samples (𝑛 ≥ 30): Let 𝑥ҧ be mean of a sample of size 𝑛 drawn from a population
having mean 𝜇 and variance 𝜎 2 , then the maximum error is

𝜎
𝐸𝑚𝑎𝑥 = 𝑍𝛼 ⋅ , where 𝐸 = |𝑥ҧ − 𝜇|
2 𝑛

𝑆 2
1 2
∗ 𝐼𝑓 𝜎 𝑖𝑠 𝑢𝑛𝑘𝑛𝑜𝑤𝑛 𝑡ℎ𝑒𝑛 𝐸𝑚𝑎𝑥 = 𝑍𝛼 ⋅ , where 𝑆 = ෍ 𝑥𝑖 − 𝑥ҧ
2 𝑛 𝑛−1
For small samples:
𝑆
𝐸𝑚𝑎𝑥 = 𝑡𝛼 ⋅
2 𝑛
iv) Confidence Interval: If 𝜃෠𝐿 and 𝜃෠𝑈 be two values such that 𝑃 𝜃෠𝐿 ≤ 𝜃 ≤ 𝜃෠𝑈 = 1 − 𝛼 then the
interval [𝜃෠𝐿 , 𝜃෠𝑈 ] is called 1 − 𝛼 100% confidence interval of the parameter 𝜃. 𝛼 is level of
significance.

Example:
1) 𝛼 = 5% = 0.05
𝛼
⇒ = 0.025
2
1 − 𝛼 100% = 95%
𝛼
𝑃 𝑧 ≥ 𝑧𝛼 = 0.025 =
2 2

𝑃 𝑧 ≤ 𝑧𝛼 = 1 − 0.025 = 0.975
2

𝐹 𝑧𝛼 = 𝐹(1.96)
2

𝑧𝛼 = 1.96.
2
2) 𝛼 = 2% = 0.02
𝛼
= 0.01
2
100 1 − 𝛼 % = 98%

𝛼
𝑃 𝑧 ≥ 𝑧𝛼 = 0.01 =
2 2

𝑃 𝑧 ≤ 𝑧𝛼 = 1 − 0.01 = 0.99
2

𝐹 𝑧𝛼 = 𝐹(2.33)
2

𝑧𝛼 = 2.33.
2
3) 𝛼 = 1% = 0.01
𝛼
= 0.005
2
100 1 − 𝛼 % = 99%

𝛼
𝑃 𝑧 ≥ 𝑧𝛼 = 0.005 =
2 2

𝑃 𝑧 ≤ 𝑧𝛼 = 1 − 0.005 = 0.995
2

𝐹 𝑧𝛼 = 𝐹(2.58)
2

𝑧𝛼 = 2.58.
2
Note:

𝜎
𝑥ҧ − 𝜇 ≤ 𝑧𝛼 ⋅ = 𝐸𝑚𝑎𝑥
2 𝑛

𝜇 − 𝑥ҧ ≤ 𝐸𝑚𝑎𝑥

−𝑧𝛼 ≤ 𝑧 ≤ 𝑧𝛼 𝑥ҧ − 𝐸𝑚𝑎𝑥 ≤ 𝜇 ≤ 𝑥ҧ + 𝐸𝑚𝑎𝑥


2 2

∴ 𝜇 ∈ [𝑥ҧ − 𝐸𝑚𝑎𝑥 , 𝑥ҧ + 𝐸𝑚𝑎𝑥 ]


𝑥ҧ − 𝜇
−𝑧𝛼 ≤ ≤ 𝑧𝛼
2 𝜎 2
𝑛
𝜎 𝜎
−𝑧𝛼 ⋅ ≤ 𝑥ҧ − 𝜇 ≤ 𝑧𝛼 ⋅
2 𝑛 2 𝑛
1) Find 𝛼, 𝑧𝛼
2
2) Find 𝛼, 𝑧𝛼
2
Testing of Hypothesis

Hypothesis: A base statement regarding population parameter which is to be tested is called


hypothesis.
Types:
1) Null Hypothesis: A hypothesis which is unbiased and having no difference is called null hypothesis,
it is denoted by Ho .
2) Alternative Hypothesis: A hypothesis which contradicts null hypothesis is called alternative
hypothesis. It is denoted by H1 .
Examples:

1) H0 : 𝜇 = 65 𝑘𝑔 2) H0 : 𝜇 ≥ 8 𝑘𝑔 3) H0 : 𝜇 ≤ 200 𝑘𝑔
H0 : 𝜇 ≠ 65 𝑘𝑔 H0 : 𝜇 < 8 𝑘𝑔 H0 : 𝜇 > 200 𝑘𝑔

(Two tailed) Left tailed Right tailed


Error Sampling:
1. Type-I error: Rejection of H0 even when it is true. i.e., Null hypothesis is true but it is rejected
by the test procedure.
Probability (type-I error) = 𝛼
where 𝛼 is called level of significance.
2. Type-II error: Accepting H0 even when it is false. i.e., Null hypothesis H0 is false but it is
accepted by the test procedure.
Probability(type-II error)=𝛽
1 − 𝛽 is called power of the test.
Critical Region: A region corresponding to test statistic in the sample space which lead to
rejection of Null hypothesis 𝐻0 is called critical region (or) rejection region.

The region in which the null hypothesis 𝐻0 is accepted is called acceptance region.

Critical Values (Significant Values): The value (values) of test statistic which divide the total
sample space into acceptance region and rejection region are called critical values or significant
values.
i) Two-tail ii) Right-tail

Critical values are: Critical value is: 𝑧𝛼


−𝑧𝛼 and 𝑧𝛼
2 2
ii) Left-tail

Critical value is:


𝑧𝛼 or 𝑧1−𝛼

You might also like