To predict the probability of tax revenue, a suitable predictive model would be a regression model. Specifically, a linear regression model or a logistic regression model could be used depending on the nature of the data and the specific requirements of the analysis.

Linear regression models are used to predict continuous variables, such as tax revenue amounts. The model would be trained on historical data that includes information about the tax revenue collected in previous periods, as well as any relevant economic, demographic, or policy variables that might affect tax revenue. The resulting model would then be used to predict future tax revenue based on these inputs.

Logistic regression models, on the other hand, are used to predict binary outcomes, such as the probability of tax revenue meeting a certain threshold. For example, if the goal is to predict the probability of tax revenue exceeding a certain amount, a logistic regression model could be trained using historical data to identify the factors that are most predictive of meeting this threshold. This model could then be used to make predictions about future tax revenue probabilities.

It's important to note that the quality of any predictive model depends on the quality and completeness of the data used to train it, as well as the accuracy and relevance of the variables included in the model. Therefore, careful attention should be paid to data cleaning, feature selection, and model validation to ensure that the model produces accurate and reliable predictions.

The best method for calculating future tax revenue is to use a micro-analytical simulation model using the anonymized individual (panel) data of taxpayers. This allows an estimation accuracy of almost 100% to be achieved.

how about performance of neural network in comparison with regression?, several variables in the dataset contain the realization of income per individual which has been grouped by land value zone