How to find out which indicators are the best to predict a trend

[tiger2014]

I have a table that the first column is a stock price sequence, and the second column is a tag that shows where the trend starts and ends, other columns are the value of many indicators (for example MACD, RSI, EMA, and so on), every indicator uses one column.

I want to use ML.NET to find out which indicators/indicators combination are most important to identify a trend (up/down trend).

I want to know if this can be resolved by ML.NET and which model I should be selected Binary classification, Multi-class classification, or others.

[tiger2014]

Terrible, nobody responded to it. I need to give it up and pick up python ML.

[Infernogates]

Did you try asking chatGPT? "Based on your problem description, you are trying to predict the direction of the stock price trend (up or down) based on the values of different indicators. This is a binary classification problem since there are only two possible outcomes (up or down).

To solve this problem, you can use a binary classification algorithm in ML.NET. You will need to split your dataset into training and testing sets and train the algorithm on the training data. Once the model is trained, you can use it to predict the direction of the trend on the testing data.

As for identifying the most important indicators or combination of indicators, you can use feature selection techniques, such as correlation analysis or mutual information analysis, to identify which features are most correlated with the target variable (trend direction). You can also use algorithms that automatically select the most relevant features, such as Lasso regression or Random Forest.

Overall, I would recommend using a binary classification algorithm and exploring feature selection techniques to identify the most important indicators for your prediction problem."

[tiger2014]

Thank you!