For prediction from long-term historical data which Machine learning algorithm would be appropriate ? Initially I am thinking about LSTM but not confident about the possible outcome.
There is no best Machine Learning (ML) concept or algorithm for tasks. The choice of ML is affected by several factors such as: Data size, type, accuracy, complexity vs linearity, number of features, etc. First, you need have a good understanding of your data and what you want from the data. The data must be cleaned. Then carry out brief research to have a good understanding of what ML has been successfully applied on similar datasets like yours. This will give a brief idea of what ML to use.
Alternatively, you can try WEKA, after loading the data, only algorithms that are suitable for data will be enabled.
WEKAS is a collection of machine learning algorithms for data mining tasks. It contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization.
For more information on WEKA check this link
https://www.cs.waikato.ac.nz/ml/weka/
However, with experience, knowledge of MLs and datasets, it is possible to guess which ML might be suitable for a particular task.