These days, the use of machine learning for modeling has become dominant but the technique requires large data. Given the peculiar data limitations usually experienced in developing countries, could there be a suitable simple alternative to these sophisticated techniques?
Ojonimi Yusuf Usman machine learning needing lot of data is misconception. Machine learning is extension of statistical models. You will need a sample to build a good model. I am not sure how much data you have but, one of the techniques that are used in machine learning is by building a model using k-hold technique. Also re sampleing and continuous iterations during training will help you build reliable models.
@Nagdev, thank you for your answer. The data I have is 25 year's annual data.
Use resampling techiniques like k-fold cross-validation with k=5 or k=10. You can use repeated train-test-split n times with n=1000 to train your model with less data.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling