TAFM: A Recommendation Algorithm Based on Text-Attention Factorization Mechanism

Table 1

Dataset field descriptions.

Field name	Description	Numeric type
m_sex	User gender	Discrete
m_access_frequencies	Frequency of user access	Discrete
m_twoA∼m_twoE	Anonymous fields for dichotomous user features	Discrete
m_categoryA∼ m_categoryE	Anonymous fields for user classification features	Discrete
m_num_interest_topic	Number of topics of interest to users	Discrete
num_topic_attention_intersection	Intersection count of user-focused topics and question-bound topics	Discrete
q_num_topic_words	Number of topics bound to an issue	Discrete
num_topic_interest_intersection	Intersection count of topics of interest to users and question-bound topics	Discrete
m_salt_score	User salt score	Continuous
m_num_atten_topic	Number of topics followed by users	Continuous
q_num_title_chars_words	Question title word count	Continuous
q_num_desc_chars_words	Number of characters in the problem description	Continuous
q_num_desc_words	Number of words in the problem description	Continuous
q_num_title_words	Number of words in the question title	Continuous
days_to_invite	Number of days since the invitation was created for the issue	Continuous