Description
Data Science in Python: An Unsupervised Course. This training course is designed using a hands-on, project-oriented approach that will help you master the basics of self-paced learning using Python. To kick off the course, looking at the data science workflow, we’ll look at methods and applications of unsupervised learning, as well as the steps to prepare data for modeling. In this section, you will learn how to set the right modeling granularity, use feature engineering techniques, select appropriate features, and scale data using standardization and normalization.
We then move on to implementing, tuning, and interpreting three popular clustering models using scikit-learn. We’ll start with K-means clustering, learn how to interpret the center of the output clusters, and use inertia plots to select the right number of clusters. Next we look at hierarchical clustering, in which a dendrogram is used to identify clusters and a cluster map is used to interpret them. Finally, we will use DBSCAN to detect clusters and noise points and evaluate models using silhouette estimation. We will also use DBSCAN and Isolation Forest for anomaly detection, which is a common use for unsupervised learning models to identify outliers and unusual patterns. You’ll learn how to set up and interpret the results of each model, and visualize anomalies using paired plots.
In the next section, we’ll introduce the concept of dimensionality reduction, discuss its benefits in data science, and look at steps in the data science workflow where the technique can be applied. We’ll then look at two popular methods: principal component analysis, which is great for both feature extraction and data visualization, and t-SNE, which is ideal for data visualization. Finally, we’ll introduce you to recommendation engines, and you’ll practice building content-based recommendation systems and collaborative filtering using techniques such as analogous cosine and singular value decomposition.
During the course, you will play the role of a junior data scientist on the HR analytics team of a software company that is trying to improve employee retention. Using the skills you’ll gain in the course, you’ll use Python to segment employees, visualize clusters, and suggest next steps to improve employee retention.
What will you learn
- Proficient in the fundamentals of unsupervised machine learning in Python, including clustering, anomaly detection, feature reduction, and recommendation systems.
- Prepare data for modeling using feature engineering, selection, and scaling.
- Implementation, tuning, and interpretation of three types of clustering algorithms: K-Means clustering, hierarchical clustering, and DBSCAN.
- Using unsupervised learning methods such as Isolation Forest and DBSCAN for anomaly detection.
- Application and interpretation of two types of feature reduction models: principal component analysis (PCA) and t-SNE.
- Building recommendation systems using content filtering and collaborative filtering techniques, including analogous cosine and singular value decomposition (SVD).
Who is this course suitable for?
- Data scientists who want to learn how to create and interpret unsupervised learning models in Python.
- Analysts or business intelligence (BI) professionals looking to learn about unsupervised learning or move into a data scientist role.
- Who cares?
Data Science in Python: Unsupervised Learning Course Specifications
- Publisher: Udemy
- Instructor: Maven Analytics , Alice Zhao
- Level of training: from beginner to advanced
- Duration of training: 16 hours 46 minutes.
- Number of courses: 202
Course headings
Data Science in Python: Unsupervised Course Prerequisites
- We highly recommend taking our Data Preparation and EDA course before doing this.
- Jupyter Notebooks (free download, we’ll walk you through installation)
- Familiarity with basic Python and Pandas is preferred, but not required.
course images
Example video course
installation instructions
Once extracted, watch using your favorite player.
Subtitles: No
Quality: 720p
Download link
Password for file(s): www.downloadly.ir
size
4.9 GB