Social media provide a low-cost alternative source for public health surveillance and health-related classification plays an important role to identify useful information. We summarized the recent classification methods using social media in public health. These methods rely on bag-of-words (BOW) model and have difficulty grasping the semantic meaning of texts. Unlike these methods, we present a word embedding based clustering method. Word embedding is one of the strongest trends in Natural Language Processing (NLP) at this moment. It learns the optimal vectors from surrounding words and the vectors can represent the semantic information of words. A tweet can be represented as a few vectors and divided into clusters of similar words. According to similarity measures of all the clusters, the tweet can then be classified as related or unrelated to a topic (e.g., influenza). Our simulations show a good performance and the best accuracy achieved was 87.1%. Moreover, the proposed method is unsupervised. It does not require labor to label training data and can be readily extended to other classification problems or other diseases.

Keywords: Big data, Word Embeddings, Word2Vec, Machine learning, Natural Language Processing, Clustering Process, unsupervised classification

The method consists of 3 steps:

Step 1: text preprocessing - Social media are informal, less structured, and contain misspellings and non-textual information.
Step 2: clustering process - This step divides a tweet into clusters of words.

Our clustering process is inspired by the following picture - the words are clustering together with topics in a vector space.

A tweet can be divided into clusters of words



Step 3: similarity measure

Cosine similarity is used to identify whether a cluster is related to a topic

Two vectors are highly similar if their cosine similarity value is approaching 1.

PDF Download Links