---
title: "Cold start in collaborative filtering"
subject: "Information System"
chapter: 7
key: "semester_8/information_system:cold-start-collaborative-filtering"
source: https://www.bctengineer.com/q/semester_8/information_system/cold-start-collaborative-filtering
site: BCT Engineering
---

# Cold start in collaborative filtering

Information System · Chapter 7: Web based information system and navigation

- Asked in: 2074 Bhadra, 5 marks; 2070 Bhadra, 3 marks (asked 2 times)
- Video: [The Cold Start Problem of Data Science](https://www.youtube.com/watch?v=F81pBEXLjC8) — Microsoft in Business
- Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14
- Page: https://www.bctengineer.com/q/semester_8/information_system/cold-start-collaborative-filtering
- Whole chapter: https://www.bctengineer.com/md/semester_8/information_system/chapter-7

## Question

What do you mean by cold start problem in collaborative filtering? Explain with example.

## Short answer

The cold start problem occurs when collaborative filtering lacks rating data for new users, new items, or a new system, preventing neighbor similarity computations. For example, a new user Gita with no ratings cannot be compared to existing users, and film D with no ratings is never recommended. Solutions include hybrid methods, initial rating prompts, and demographic data.

## Answer

#### Collaborative filtering (CF)
**Collaborative filtering** recommends items to a user based on the ratings and behaviour of other similar users (or items). It needs a user-item rating matrix.

#### Cold start problem
The **cold start problem** is the difficulty of giving good recommendations when the system has little or no data about a new user, a new item or a new system. Because CF depends on past ratings, it cannot find similar users or items without them.

#### Types
1. **New user problem:** a user with no ratings has no neighbours, so no recommendation can be computed.
2. **New item problem:** a new item has no ratings, so it is never recommended; the more popular items get all the attention.
3. **New system (new community) problem:** at the start, there are no ratings at all.

#### Example
A movie website has users Ram, Sita and Hari who rated many films.
| | Film A | Film B | Film C |
|---|---|---|---|
| Ram | 5 | 4 | 1 |
| Sita | 5 | 5 | 2 |
| New user Gita | ? | ? | ? |
| New film D | no ratings | | |

CF can compare Ram and Sita because they have rated common films. It cannot compare Gita (no ratings) with anyone. Film D will not be recommended to anyone because nobody has rated it.

#### Solutions
- Ask new users to rate a few items or choose genres on sign-up.
- Use demographic data (age, location) or the user's social-media information.
- Use content-based filtering for new items (genre, actor, description).
- Hybrid methods that combine content-based and CF.
- Recommend popular or trending items at first.
- Let experts rate new items; use active learning.

## Check yourself

The answer's key points as multiple-choice questions, with the correct option and why.

1. What is the cold start problem in collaborative filtering?
   A. Difficulty giving good recommendations when there is little or no data about a new user, item or system
   B. Slow processing because the server is overheated
   C. Too many ratings for every item
   D. Users giving only negative ratings
   Answer: A. CF needs past ratings; without them it cannot find similar users or items. (Key point: Meaning)

2. Why can Gita, a new user with no ratings, not get CF recommendations?
   A. Her name is not in the film list
   B. She rated every film the same
   C. The film list is too short
   D. She has no ratings, so she cannot be compared with other users to find neighbours
   Answer: D. Similarity is computed from common ratings, which she lacks. (Key point: New user)

3. What happens to new film D that nobody has rated?
   A. It is recommended to everyone
   B. It is ranked first as the newest item
   C. It is never recommended because it has no ratings
   D. It is rated automatically by the system
   Answer: C. CF depends on ratings, so unrated items are ignored. (Key point: New item)

4. Which is a solution to cold start?
   A. Delete all existing ratings
   B. Use content-based filtering or ask new users to rate a few items
   C. Stop recommending any items
   D. Recommend only items the user already bought
   Answer: B. Content features, onboarding questions, demographics or popularity can fill the gap. (Key point: Solution)

## Related in Information System

- [Collaborative filtering concept](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Acollaborative-filtering-concept) — Information System, chapter 7
- [Short note: collaborative filtering](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Ashort-note-collaborative-filtering) — Information System, chapter 7

## Related in other subjects

- [Content filtering vs packet filtering](https://www.bctengineer.com/question.md?key=semester_8%2Finternet_and_intranet%3Acontent-filtering-packet-filtering) — Internet and Intranet, chapter 5