Python package for building data embeddings
Project description
build2Vec
Graph Neural Networks based building representation in the vector space
Installation
$ pip install build2vec
Examples
import networkx as nx
from build2vec import Build2Vec
emb_dimensions = 10
# Create a graph using networkx -- you can generate the graph from dataframe of edges
graph = nx.from_pandas_edgelist(df_links_graph)
build2vec = Build2Vec(graph, dimensions=emb_dimensions, walk_length=50, num_walks=50, workers=1)
model = build2vec.fit(window=50, min_count=1, batch_words=10)
Todos:
- Add automatic grid generation method.
- Add automatic graph construction method.
- Add visualization moddule.
- Add ML clustering, classification, and prediction moduels.
- Define other builing-related random walks methods.
Citation:
@inproceedings{abdelrahmanbuild2vec,
title = {{Build2Vec: Building Representation in Vector Space}},
year = {2020},
booktitle = {SimAUD 2020},
author = {Abdelrahman, Mahmoud M and Chong, Adrian and Miller, Clayton},
number = {May},
pages = {101--104},
publisher = {Society for Modeling {\&} Simulation International (SCS)},
url = {http://simaud.org/2020/proceedings/102.pdf},
address = {Online},
arxivId = {2007.00740},
keywords = {Feature learning, Graph embeddings, Representation learning, STAR, node2vec}
}
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
build2vec-0.0.5.tar.gz
(5.5 kB
view hashes)
Built Distribution
Close
Hashes for build2vec-0.0.5-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 806196b6fc17a2a5fd637714f3b1bf775b7f05dc87b894ad063f7260f46775c8 |
|
MD5 | 4580553da67ad528cc276ff577cd931b |
|
BLAKE2b-256 | 7bfa823625500075def07d9d2e88d90e4ad53442d747c4f07f0d37e4b672ac68 |