Unsupervised Learning - Kmeans Week 11 Challenge

Q1.

1. Hamming Distance: Simple and effective for binary or nominal categorical data.
2. Simple Matching Coefficient (SMC): Good for symmetric data where matching and mismatching attributes are equally important.
3. Jaccard/Dice Similarity: Preferable for sparse data (e.g., binary data with many 0's) or when you only care about the "positive" (1) matches.
4. Gower's Similarity: Best for mixed datasets (both categorical and numerical features).
5. One-Hot Encoding + Euclidean: Useful when you want to treat categorical data similarly to numerical data and combine it with traditional distance measures.

---

---

Q2.

Note: Empty rows were there in provided dataset so manually deleted rows in excel

imported libraries and dataset.

Filled missing values in cylinders

K-Mean is performed to get clusters and with that business insight is given

Business Insight

Cluster 0: Fuel-Efficient Compact Cars.

• This cluster likely represents customers interested in fuel-efficient, compact cars ideal for city driving. They prioritize cost-effectiveness, low maintenance, and higher fuel economy.

Cluster 1: Family-Friendly Mid-Range Vehicles

• This cluster represents families or customers who seek a balance between performance, space, and fuel efficiency. Cars in this group might be mid-range sedans, crossovers, or SUVs that offer comfort and versatility

Cluster 2: High-Performance or Luxury Vehicles

• This cluster is characterized by high-performance, luxury cars. These cars have powerful engines and lower fuel economy, appealing to customers who value performance, luxury, and brand prestige.

Product Positioning and Development:

The clusters reveal distinct segments that can guide product positioning. Each cluster represents a different market segment with unique preferences and needs.

Optimizing Inventory Management:

Clusters provide information on which types of cars are likely to be in higher demand in different regions or dealerships.

Tailored Marketing Strategies:

Different clusters allow for tailored marketing campaigns. Each cluster corresponds to a unique consumer profile with specific needs and preferences.

Price Differentiation Strategy:

Understanding the cluster characteristics can help in setting a price strategy. Cars in the high-performance or luxury cluster can be priced at a premium, while cars in the economy cluster can be competitively priced.