emre şahin's digital garden 🍃 - graphs https://emresahin.net/tags/graphs/ Posts in the graphs tag en contact@emresahin.net (Emre Şahin) Tue, 15 Sep 2026 19:46:32 +0000 Graphs for Recommendation Systems 2024-02-23T19:36:41+00:00 2024-02-23T19:36:41+00:00 Emre Şahin Fri, 23 Feb 2024 19:36:41 +0000 https://emresahin.net/graphs-for-recommendation-systems/ https://emresahin.net/graphs-for-recommendation-systems/ This is an older note: I think using graphs for recommendation systems is not a natural choice where temporal patterns and sequences play a role. It is not natural to represent time with nodes that can have multiple edges. Instead, RNNs or Markov chains might be more suitable. Recommendation is a... machine-learning data-science graphs recommendation-systems sequential-modeling This is an older note:

I think using graphs for recommendation systems is not a natural choice where temporal patterns and sequences play a role. It is not natural to represent time with nodes that can have multiple edges. Instead, RNNs or Markov chains might be more suitable. Recommendation is almost always sequential. Otherwise, it is about finding embeddings between elements. If so, graphs might not fit these problems. If graphs are not suitable, why do we use them?

The answer is simple: People don’t buy things in a strict sequence; they have options, multiple choices, and various factors that play a role in these choices. So, it is actually natural to model these interactions with graphs.

]]> Turning Ottoman Letters into Graphs (1) 2012-09-22T17:00:00+00:00 2012-09-22T17:00:00+00:00 Emre Şahin Sat, 22 Sep 2012 17:00:00 +0000 https://emresahin.net/12123-4-1014/ https://emresahin.net/12123-4-1014/ Today’s work was about sharding a page’s components and recording them as new images. Instead of artificial boundaries (like word/sentence boundaries), the labeling should rely on connected components. There are two problems here. In Arabic-based writing systems, dots play a significant role, muc... Dervaze Computer Vision Document Analysis Ottoman Arabic connected components character recognition document processing graphs Today’s work was about sharding a page’s components and recording them as new images. Instead of artificial boundaries (like word/sentence boundaries), the labeling should rely on connected components.

There are two problems here. In Arabic-based writing systems, dots play a significant role, much more so than in Latin-based scripts. Therefore, these dots should be classified correctly.

The second problem is that the connected components are not always reliable. There are unduly divided components which are part of a single component. We can’t label them as they are, and uniting them into a uniform component requires manual intervention—something we try to avoid.

In the coming days, I’ll try to exemplify these problems and how we treat them.

]]>