# A guide to basic errors in machine learning

Types of machine learning models are obtained by testing them. We use a lot of statistical tests, but one thing we all know is that no statistical test is perfect. Some errors in the models are easy to understand, but difficult to obtain. The basic criterion error can be considered an easy error to understand, but a difficult one to find. The concept of basic norm error is derived from behavioral science. In this article, we will discuss this error and we will also understand its application to machine learning. The main points discussed in the article are listed below.

## Contents

1. What is the basic rate?
2. What is the main criterion error?
3. Basic speed error in machine learning
4. Why does the inaccuracy of the basic norm occur?
5. How can basic standard errors be avoided?

Let’s start by understanding the basic norm.

## What is the basic rate?

In statistics, the basic criterion can be considered as the probability of classes that are without conditional proof of characteristics. We can also think of the basic criterion as a preconceived notion. We can understand it by the example of the world’s engineers. So if 2% of the people in this world are engineers, then the basic level of engineers is only 1%.

In many statistical analyzes, we see that the basic criterion is difficult to compare. Assume that 2,000 people beat Kovid-19 using any treatment. As long as we don’t look at the entire population that has undergone such treatment, this seems like a good number. Assume that we understand that the basic rate of treatment success is only 1/50, i.e., only 2,000 people are successful in defeating the covid using the treatment, while it applies to 100,000 people. This is an important number and so we get a more accurate report on treatment using the baseline.

With the example above, we can understand how important basic information is when conducting a statistical analysis. The use of the basic criterion in statistical analysis can be called the error of the basic criterion. Let’s see what the base rate error is.

Are you looking for a complete repository of Python libraries used in data science? check here.

## What is the main criterion error?

In a general sense, we can say that error can be defined as the use of incorrect evidence, incorrect actions, or invalid actions when constructing evidence. We can say that it seems stronger than its real strength.

The error of the basic norm is also a kind of error, which is also called as bias of the basic norm and neglect of the basic norm. This type of error contains information about the base rate and specific data. Ignorance of the basic norm information exists in the interest of the individual. We can also view the error as part of a broader neglect.

## Basic speed error in machine learning

Above, we discussed that this error is something related to data override, and we know about machine learning that the models under it work on the basis of data (we can also say that data is data). Let us take the example of classification models, in which we use a misunderstanding matrix to describe the performance of classification models.

The process of constructing an error matrix takes place after testing the model in the test data, and the misunderstanding matrix informs us of the number of correct and incorrect predictions from the model. In the confusion matrix, the false-negative paradox and the false-positive paradox are examples of errors at the basic level.

Let’s say there is a machine learning model for recognizing the faces of happy people, which results in more false test results than true positive ones. We want the model to predict 99% accuracy and analyze 1000 people every day, if we evaluate it by the number of tests, then higher accuracy can be preferred and the final result will determine the false positives compared to the true ones.

We can measure the probability of a positive result based on the accuracy of the test and the quality of the selected population. In conclusion, we can say that if the part given under any condition is lower, the rate of false positives will be greater than false positives if the base rate error is present.

Let’s illustrate this with an example where the model is used to classify a population of 1,000 samples, the model says that 40% are class A and provide a false positive rate of 5% and a false-negative criterion of zero.

From Class A and positive examples

1000 X (40/100) = 400, these patterns get a true positive

Class B and negative patterns

1000 X [(100 – 40)/100] X 0.05 = 30, These samples receive a false positive

So 1000 – (400 + 30) = 570 samples are negative

The final size will be accurate

400 / (30 + 400) = 93%

The misunderstanding matrix will be:

Assume that it is applied in 1000 different samples, which is only 2% of the Class A sample, then the misunderstanding matrix will be as follows.

In this case, we can say that out of 69 samples, 20 of them are correctly predicted. Thus, the probability of a correct prediction of the model for a similar test will be 29%, which is 93% accurate.

## Why does the base rate error occur?

In the study, we can understand a number of reasons for the existence of the error, and all of them are related to a matter of importance, which says that we ignore the basic normative data. Often the base rate data is classified as irrelevant and ignores its pre-processing. Sometimes we also learn that representative heuristics are the cause of the fundamental criterion error.

## How can basic standard errors be avoided?

As discussed above, ignoring the base rate data leads to the base rate error, and we can also avoid base level errors by focusing on the base rate data. We also need to understand what examples there are that are not reliable predictions because we think about them.

When measuring the probability of an event, we are required to do more. Bayesian methods help us in measuring the probability distribution of elevations and are a way to reduce base-level errors.

## The final words

In this paper, we have examined the basic criterion errors that result from the models being used when making predictions and are due to omissions from the basic criterion data. At the same time, we discussed how this error occurs and how we can prevent it.