In data science, hypothesis testing has several functions. One is to assist in making decisions based on objective statistical evidence. Hypothesis comes from Greek, which consists of two words, namely "hupo" which means weak or lacking while "thesis" which means theory or statement presented as evidence. Thus, a hypothesis can be interpreted as a statement whose truth still needs to be proven or is a provisional conjecture.

Hypothesis Testing

Hypothesis Test is a statistical test that is useful for validating a statement or assumption that has been defined. In research we need to formulate a problem, from the formulation of the problem will arise assumptions or hypotheses.

We need to test these hypotheses or assumptions, whether the hypothesis is true or not. The main purpose of hypothesis testing is to avoid coincidence or bias from the research we take.

A. Stages of Conducting Hypothesis Test

  1. Defining hypotheses
  2. Choose the appropriate test.
  3. Calculate statistical values and determine error values.
  4. Drawing Conclusions

Case in point: Dunia Fantasi Management (DUFAN) increased the intensity of One-way ticket checks compared to previous years. This more intensive ticket inspection policy is believed to have a positive impact on DUFAN's income.

To prove this point, the initial hypothesis is:

Initial Hypothesis: THERE IS NO DIFFERENCE IN income AFTER or BEFORE changes to the ticket inspection system.

The management hopes that this hypothesis is rejected, thus proving that their opinion is correct!

Untitled

Untitled

Untitled

  1. Two-way hypothesis

    The two-way hypothesis is a hypothesis that is able to move in both negative and positive directions. Testing on hypotheses of this type is aimed at knowing its direction.

    Example: It is suspected that there is an influence between the quality of the road and the speed of the car.

  2. One-way hypothesis

    The two-way hypothesis is a hypothesis that is able to move only one way, whether in a negative or positive direction. Testing on this type of hypothesis aims to prove whether the hypothesis is true or not and test the difference or relationship between the two variables.

    Example: It is suspected that there is a positive influence between the quality of the road and the speed of the car.