Medicament latent adverse reaction discovery method based on neural network language model

Abstract

A medicament latent adverse reaction discovery method based on a neural network language model relates to the field of a medicament latent adverse reaction discovery method. The method comprises the following steps of performing data acquisition and cleaning; performing model optimization, modifying an original Skip-gram algorithm for performing characteristic extraction from the AERS report of the FDA and the DrugBank DDI dataset; extending an interaction database, selecting 5 adverse reactions kinds: kidney damage, heart toxicity, liver toxicity, blood pressure abnormity and neurotoxicity, using the five adverse reaction kinds as a Logistic regression verifying medicament and adverse reaction vector range and extending a DrugBank medicament interaction database in the range of the five kinds; performing Logistic regression and verifying a vector effect, finishing CM-TF-IDF model construction and distributed vector generation by means of Scikit-learn. The medicament latent adverse reaction discovery method based on the neural network language model settles the following problems in an existing method: relatively low data processing speed, unsuitability for processing large-scale data set and incapability of quickly performing characteristic learning.

Description

technical field [0001] The invention relates to the field of methods for discovering potential adverse drug reactions, and is a method for discovering potential adverse drug reactions based on a neural network language model. Background technique [0002] Adverse drug reaction events have always been the focus of health and medical institutions around the world. Drug clinical trials are carried out in a small range of people and special groups of people. The number of test personnel and the differences in the population cannot cover the market well. Pharmaceutical Audience. At the same time, due to the short trial period, all adverse reactions of the drug cannot be fully revealed in the clinical trial stage. The flow of drugs with unknown potential adverse drug reactions to the market will pose a threat to public health. The FDA receives reports on suspected adverse drug reaction events and mandatory reports from drug manufacturers from healthcare professionals and consume...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies of the prior art, the present invention provides a method for discovering potential adverse drug reactions based on a neural network language model, which solves the problem that the existing method for discovering potential adverse drug reactions has a slow data processing speed and is not suitable for processing large-scale data sets , and the problem of not being able to quickly perform feature learning

Images