Applied Machine Learning II
Make Data-Based Predictions with Applied Machine Learning
In Machine Learning or statistical learning, it is possible to train a computer to perform a specific task. This is extremely powerful specifically when a lot of data related to this task is available. Machine Learning is a must-have for a data scientist to go beyond data analytics.
- Ability to reduce the number of features when it is too large
- Knowing how to cluster similar records in an unlabelled dataset
- Ability to get better performance results using multiple models instead of one
Who should attend:
Applied Machine Learning I
Enterprise Data Scientist (EDS)
Applied Machine Learning II is one of the modules under our Enterprise Data Scientist (EDS) programme. EDS is a 42- day training program that provides participants with the tools to be key leaders and contributors of a data science team and be able to analyze data to drive informed business decisions.
Details of Subject
- Supervised Machine Learning: Classification –
Classification is the sub-field of Machine Learning that consists in building a model from training data (data with correct class) in order to predict the class for other data. In classification, there is only a finite number of classes
- Naïve Bayes. Algorithms of this class are generative: they assume that the data was generated by some (naïve) distribution. Which can be binomial, Gaussian, multinomial… The training part consists in finding the parameters of the distribution. The predicting part consists in finding the most likely class according to the different distributions
- Ensemble methods: Random forests. Decision trees (covered ML1) have the drawback of overfitting. To overcome this problem, in Random Forest many very simple trees are built.
- Clustering –
Clustering consists in grouping elements without any prior information. Kmean is a clustering algorithm based on the distance between the points
- Dimensionality reduction –
Dimensionality reduction consists in reducing the number of dimensions! In PCA, the goal is to find new dimensions that best explain the data
- PCA (Principal Component Analysis)
- Introduction to deep learning: neural networks – With Deep Learning, researchers manage to get very good results on some machine learning tasks. One of them is image classification.Deep learning is based on neural networks. In this class we will teach what is a neuron in a neural network and how neural networks manage to handle non-linear problem
EDS CADS Certified Enterprise Data Scientist
Certification information for this module & track will be made available soon.
Hear from Our Alumni
Register Interest for Group or Organisation Enrolment
Applied Machine Learning II