What are the missing value filling methods in spss?

spss missing value filling methods include: 1. Mean interpolation, using the mode of the attribute to fill in the missing values; 2. Using similar mean interpolation; 3. Maximum likelihood estimation, through observation The marginal distribution of data can perform maximum likelihood estimation of unknown parameters; 4. Multiple interpolation, select the most appropriate interpolation value based on a certain selection basis.

The operating environment of this tutorial: Windows 7 system, SPSS version 26.0, Dell G3 computer.

1. Mean interpolation. The attributes of data are divided into fixed-range and non-fixed-range types. If the missing values ​​are fixed-range, use the average of the existing values ​​of the attribute to interpolate the missing values; if the missing values ​​are non-fixed-range, use the mode of the attribute based on the mode principle in statistics. (i.e. the value with the highest frequency) to fill in the missing values.

2. Use similar mean interpolation. The same mean interpolation method belongs to single-value interpolation. The difference is that it uses a hierarchical clustering model to predict the type of missing variables, and then interpolates with the mean of that type. Assume that X = (X1, X2...Xp) is a variable with complete information, and Y is a variable with missing values.

Then first cluster X or its subset, and then interpolate the means of different classes according to the class to which the missing cases belong. If the introduced explanatory variables and Y need to be analyzed in future statistical analysis, then this interpolation method will introduce autocorrelation into the model, causing obstacles to the analysis.

3. Maximum likelihood estimation (Max Likelihood, ML). Under the condition that the missing type is missing at random, assuming that the model is correct for the complete sample, the unknown parameters can be estimated with maximum likelihood through the marginal distribution of the observed data (Little and Rubin).

This method is also called maximum likelihood estimation that ignores missing values. The calculation method often used in practice for maximum likelihood parameter estimation is Expectation Maximization (EM).

4. Multiple Imputation (MI). The idea of ​​multi-value interpolation comes from Bayesian estimation, which believes that the value to be interpolated is random and its value comes from the observed value. In specific practice, the value to be interpolated is usually estimated, and then different noises are added to form multiple sets of optional interpolation values. Select the most appropriate interpolation value based on a certain selection basis.

Extended information

There are many reasons for missing values. Equipment failure, inability to obtain information, inconsistency with other fields, historical reasons, etc. may produce missing values. . A typical processing method is interpolation, and the data after interpolation can be regarded as obeying a specific probability distribution. In addition, you can also delete all records containing missing values, but this operation also changes the distribution characteristics of the original data from the side.

The processing of missing values ​​is generally divided into deleting cases with missing values ​​and missing value interpolation. For subjective data, people will affect the authenticity of the data. The true values ​​of other attributes of samples with missing values ​​cannot be guaranteed, so interpolation relying on these attribute values ​​is also unreliable, so interpolation is generally not recommended for subjective data. method. Interpolation is mainly aimed at objective data, and its reliability is guaranteed.

The above is the detailed content of What are the missing value filling methods in spss?. For more information, please follow other related articles on the PHP Chinese website!