-
Notifications
You must be signed in to change notification settings - Fork 0
swapdewalkar/Triplet-Generation
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
1. PreProcessing: 1.a (vir) python process_main.py [INPUT File in Data Folder] [SPECIFY OUTPUT FOLDER (autocreate in Results/)] [PARAMETERS] - e.g. python process_main.py reviews_small.txt reviews_small_output_test PDTOWC 2. Create Datasets: (openke) 2.a python make_dataset.py [ INPUT FOLDER PATH ] [INPUT FOLDER] - Output Folder will be created in OpenKE/Dataset depending on input Folder - Two Folder will be create: With and Without Co-occurence - e.g. python make_dataset.py ~/code/Final\ Code/Result Reviews_NEW_OUTPUT_11 - e.g. python make_dataset.py ~/code/Final\ Code/Result Reviews_V10K 2.b python analysis.py Dataset/[FOLDER] - Count Relation-wise, Head-wise and Tail-wise. - create file analysis.txt for Results - e.g. python analysis.py Dataset/Reviews_V10K_WN_DP > Dataset/Reviews_V10K_WN_DP/analysis.txt 2.c python graph_analysis.py Dataset/[FOLDER] - Relation and Entity Sorted by Vocab, Count and ID and Find unique count. - entity/relation/train_stats.txt will be created - e.g.python graph_analysis.py Dataset/Reviews_V10K_WN_DP 2.d python remove_duplicate.py Dataset/[FOLDER] - Create Duplicated and Unique limited relation. - Created folder Dataset/[FOLDER]_Unique - After this you can run analysis.py and graph_analysis.py for more info - e.g. python remove_duplicate.py Dataset/Reviews_V10K_WN_DP 2.e Useful DP - python remove_dp_relations.py [Folder] - remove_dp_relations.py ===>>> useful relation. 3. Training: (openke) 3.a python train.py python train.py [INPUT FOLDER in Dataset] [OUTPUT FOLDER in res] [loss_file_modifier] - python train.py Reviews_NEW_OUTPUT_11_WN_DP Reviews_OUT_Adadelta_E1000_L1_D150_B32 0 4. Evaluation: 4.c (openke) python convert_vector.py [Dataset] [FOLDER] - python convert_vector.py Reviews_All_Rels Reviews_Adadelta_E5000_B32_D150_L1 - create embedding vector for processing 4.b (vir) python evaluate.py [path to embedding vector] [output_modifier] - Evaluate learnt embedding based on various Evaluation. - create file OUT_out_modifier
About
Triplet generation from text data using Dependency Parse Tree , WordNet and Co-Occurence.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published