Book section Open Access

Query reformulation based on word embeddings: A comparative study

Panos Panagiotou; George Kalpakis; Theodora Tsikrika; Stefanos Vrochidis; Ioannis Kompatsiaris


JSON Export

{
  "files": [
    {
      "links": {
        "self": "https://zenodo.org/api/files/076db14d-859a-4a26-ae04-4e5d8003550c/Query%20reformulation%20based%20on%20word%20embeddings.pdf"
      }, 
      "checksum": "md5:d1a1b274a55c547cc112a79775f08162", 
      "bucket": "076db14d-859a-4a26-ae04-4e5d8003550c", 
      "key": "Query reformulation based on word embeddings.pdf", 
      "type": "pdf", 
      "size": 562985
    }
  ], 
  "owners": [
    45413
  ], 
  "doi": "10.5281/zenodo.3947769", 
  "stats": {
    "version_unique_downloads": 205.0, 
    "unique_views": 201.0, 
    "views": 228.0, 
    "version_views": 228.0, 
    "unique_downloads": 205.0, 
    "version_unique_views": 201.0, 
    "volume": 117663865.0, 
    "version_downloads": 209.0, 
    "downloads": 209.0, 
    "version_volume": 117663865.0
  }, 
  "links": {
    "doi": "https://doi.org/10.5281/zenodo.3947769", 
    "conceptdoi": "https://doi.org/10.5281/zenodo.3947768", 
    "bucket": "https://zenodo.org/api/files/076db14d-859a-4a26-ae04-4e5d8003550c", 
    "conceptbadge": "https://zenodo.org/badge/doi/10.5281/zenodo.3947768.svg", 
    "html": "https://zenodo.org/record/3947769", 
    "latest_html": "https://zenodo.org/record/3947769", 
    "badge": "https://zenodo.org/badge/doi/10.5281/zenodo.3947769.svg", 
    "latest": "https://zenodo.org/api/records/3947769"
  }, 
  "conceptdoi": "10.5281/zenodo.3947768", 
  "created": "2020-07-16T09:01:30.371029+00:00", 
  "updated": "2021-02-19T12:05:07.440493+00:00", 
  "conceptrecid": "3947768", 
  "revision": 4, 
  "id": 3947769, 
  "metadata": {
    "access_right_category": "success", 
    "part_of": {
      "title": "Technology Advances and Support for Security Practitioners"
    }, 
    "doi": "10.5281/zenodo.3947769", 
    "version": "1.0", 
    "language": "eng", 
    "title": "Query reformulation based on word embeddings: A comparative study", 
    "license": {
      "id": "CC-BY-4.0"
    }, 
    "related_identifiers": [
      {
        "scheme": "doi", 
        "identifier": "10.5281/zenodo.3947768", 
        "relation": "isVersionOf"
      }
    ], 
    "relations": {
      "version": [
        {
          "count": 1, 
          "index": 0, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "3947768"
          }, 
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "3947769"
          }
        }
      ]
    }, 
    "imprint": {
      "publisher": "Springer"
    }, 
    "communities": [
      {
        "id": "connexions-h2020"
      }
    ], 
    "grants": [
      {
        "code": "700024", 
        "links": {
          "self": "https://zenodo.org/api/grants/10.13039/501100000780::700024"
        }, 
        "title": "Retrieval and Analysis of Heterogeneous Online Content for Terrorist Activity Recognition", 
        "acronym": "TENSOR", 
        "program": "H2020", 
        "funder": {
          "doi": "10.13039/501100000780", 
          "acronyms": [], 
          "name": "European Commission", 
          "links": {
            "self": "https://zenodo.org/api/funders/10.13039/501100000780"
          }
        }
      }, 
      {
        "code": "786731", 
        "links": {
          "self": "https://zenodo.org/api/grants/10.13039/501100000780::786731"
        }, 
        "title": "InterCONnected NEXt-Generation Immersive IoT Platform of Crime and Terrorism DetectiON, PredictiON, InvestigatiON, and PreventiON Services", 
        "acronym": "CONNEXIONs", 
        "program": "H2020", 
        "funder": {
          "doi": "10.13039/501100000780", 
          "acronyms": [], 
          "name": "European Commission", 
          "links": {
            "self": "https://zenodo.org/api/funders/10.13039/501100000780"
          }
        }
      }
    ], 
    "keywords": [
      "query expansion", 
      "word embeddings", 
      "terrorism"
    ], 
    "publication_date": "2020-04-30", 
    "creators": [
      {
        "affiliation": "Information Technologies Institute, Centre for Research and Technology Hellas", 
        "name": "Panos Panagiotou"
      }, 
      {
        "affiliation": "Information Technologies Institute, Centre for Research and Technology Hellas", 
        "name": "George Kalpakis"
      }, 
      {
        "affiliation": "Information Technologies Institute, Centre for Research and Technology Hellas", 
        "name": "Theodora Tsikrika"
      }, 
      {
        "affiliation": "Information Technologies Institute, Centre for Research and Technology Hellas", 
        "name": "Stefanos Vrochidis"
      }, 
      {
        "affiliation": "Information Technologies Institute, Centre for Research and Technology Hellas", 
        "name": "Ioannis Kompatsiaris"
      }
    ], 
    "access_right": "open", 
    "resource_type": {
      "subtype": "section", 
      "type": "publication", 
      "title": "Book section"
    }, 
    "description": "<p>Formulating effective queries for retrieving domain-specific content from the Web and social media is very important for practitioners in several fields, including law enforcement analysts involved in terrorism-related investigations. Query reformulation aims at transforming the original query in such a way, so as to increase the search effectiveness by addressing the vocabulary mismatch problem. This work presents a study comparing the performance of global versus local word embeddings models when applied for query expansion. Two query expansions methods are employed (i.e., CombSum and Centroid) for defining the most similar terms to each query term, based on Glove pre-trained global embeddings and local models trained on four large-scale benchmark and one terrorism-related datasets. We assessed the performance of the global and local models on the benchmark datasets based on commonly used evaluation metrics, and performed a qualitative evaluation of the respective models on the terrorism-related dataset. Our findings indicate that the local models yield promising results on all datasets.</p>"
  }
}
228
209
views
downloads
All versions This version
Views 228228
Downloads 209209
Data volume 117.7 MB117.7 MB
Unique views 201201
Unique downloads 205205

Share

Cite as