BioMedicEngine

BioMedicEngine is a vector-model based search engine over a BioMedical document collection.

Manos Chatzakis (emmanouil.chatzakis@epfl.ch)

How to run

BioMedicEngine is a command line application.

Index Creation

To index a medical file directory, use:

mkdir ${COLLECTION_OUTPUT_DIRECTORY}partialIndexing
java -jar $EXEJAR_NAME -mode indexer -input $COLLECTION_INPUT_DIRECTORY -output $COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

Query Answering

For simple query answering using the vector model implementation, run:

java -jar $EXEJAR_NAME -mode retriever -collection $BIG_COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

For medical-type weighted query answering, run:

java -jar $EXEJAR_NAME -mode topicRetriever -collection $BIG_COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

Run Automation

The following bash script file is provided to automate the run, located in BiomedicEngine/bash/run.sh:

EXEJAR_NAME=../target/BioMedicEngine-1.0-SNAPSHOT-exejar.jar
COLLECTION_INPUT_DIRECTORY=../sample/
COLLECTION_OUTPUT_DIRECTORY=/mnt/c/Users/manos/Desktop/simple_example/
BIG_COLLECTION_OUTPUT_DIRECTORY=/mnt/c/BioMedicIndexer_2/
GR_STOPWORDS=../stopwords/stopwordsGr.txt
EN_STOPWORDS=../stopwords/stopwordsEn.txt

mkdir ${COLLECTION_OUTPUT_DIRECTORY}partialIndexing
java -jar $EXEJAR_NAME -mode indexer -input $COLLECTION_INPUT_DIRECTORY -output $COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

#java -jar $EXEJAR_NAME -mode retriever -collection $BIG_COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

java -jar $EXEJAR_NAME -mode topicRetriever -collection $BIG_COLLECTION_OUTPUT_DIRECTORY -gr $GR_STOPWORDS -en $EN_STOPWORDS

Experimental Evaluation

The experimental evaluation contacted on a machine of 8 cores and 16GB memory. The results are presented in the pdf report file demonstrating query answering times for a selected collection of queries, accuracy measuserements based on BPREF and other performance evaluation components, such as index creation time and memory usage.

About

This project was the final assignment of the undergraduate/graduate course cs463-Information Retrieval Systems, taught at Computer Science Department of University of Crete.