How To Choose The Best Machine Learning Algorithms and Tools
We need to know that Machine Learning is not far from what is known as the Industrial Revolution 4.0 era. Machine Learning can be applied in all fields in this industry to help the development of the industrial world so that it is increasing and can fulfill and be helpful for the lives of people in all corners of the world.
Machine Learning has several algorithms in general, such as:
- Supervised Learning: The system must be given training in advance with factual data. Then the system will learn and form existing data patterns to be used in the prediction process. This learning process is like students being taught by their teacher.
- Unsupervised Learning: The system can rely on data that has not been trained before and can form descriptive data patterns, not predictions.
- Reinforced and Deep Learning: Learning systems from environmental feedback with learning techniques that are iterative (repetitive) and adaptive (adjust), like how humans learn by themselves. Algorithms force the learning system to find optimal values by trial and error.
Read also : Characteristic of Big Data
The next question is how do we choose the correct algorithm with the right number of existing algorithms?
How to choose can with the following tips:
- Determine the type of problem.
- The size of the training data.
- Accuracy and data processing time.
- Linearity of the algorithm.
- Number of parameters.
- Number of features
Then what about the tools we need to use to conduct the analysis?
Read also : Data Science in E-commerce
We must know that “Choosing tools should not be a major problem. Use one that is available and gets the job done quickly.”
So, if you still need to conduct experiments using code, then we can use tools that we can use more quickly to try and error first (tools that click click click or just drag and drop). Such as open source tools: WEKA, KNIME, RapidMiner, and even Ms. We can also use EXCEL to do the initial stage of analysis.
After that, we can use other tools.
