sim's Introduction

Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction

Implementation of Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction using tensorflow

Prerequisites

Python 2.x
Tensorflow 1.15.0

Data

Amazon Book Data

Getting Started

First we need to prepare data.

Amazon Prepare

Because getting and processing the data is time consuming，we had processed Amazon data and upload it for you.

tar -xzf data.tar.gz

Running

usage: train_taobao_and_book.py [-h] [-mode MODE] [-seed SEED]
                                [-use_first_att USE_FIRST_ATT]
                                [-first_att_top_k FIRST_ATT_TOP_K]
                                [-use_vec_loss USE_VEC_LOSS]
                                [-long_seq_split LONG_SEQ_SPLIT]
                                [-short_seq_split SHORT_SEQ_SPLIT]
                                [-short_model_type SHORT_MODEL_TYPE]
                                [-long_model_type LONG_MODEL_TYPE]
                                [-save_iter SAVE_ITER]
                                [-test_iter TEST_ITER] [-max_len MAX_LEN]
                                [-seq_len SEQ_LEN]
                                [-epoch EPOCH] [-memory_size MEMORY_SIZE]
                                [-batch_size BATCH_SIZE]
                                [-search_mode SEARCH_MODE] [-level LEVEL]
                                [-data_type DATA_TYPE]
                                [-att_func ATT_FUNC]

Base Model

The example for DNN

python train.py -mode train \
-data_type book \
 -max_len 100 \
 -short_model_type DIN \
 -short_seq_split '90:100' \
 -long_model_type DNN \
 -long_seq_split '0:90' \
 -seed 2  \
  -epoch 2 \
 -save_iter 10 \
 -test_iter 20 \
 -search_mode 'None'

The model type below had been supported:

DNN
DIN
MIMN

SIM

You can train SIM with two kinds of search unit:

hard-search

python train.py -mode train \
-data_type book \
 -max_len 100 \
 -short_model_type DIN \
 -short_seq_split '90:100' \
 -long_model_type DIN \
 -long_seq_split '0:90' \
 -seed 2  \
  -epoch 2 \
 -save_iter 10 \
 -test_iter 20 \
 -search_mode 'cate' \
 -att_func 'dot'

soft-search

python train.py -mode train \
-data_type book \
 -max_len 100 \
 -short_model_type DIN \
 -short_seq_split '90:100' \
 -long_model_type DIN \
 -long_seq_split '0:90' \
 -seed 2  \
  -epoch 2 \
 -data_thread_num 5 \
 -save_iter 10 \
 -test_iter 20 \
 -search_mode 'None' \
  -use_first_att True \
 -first_att_top_k 50 \
 -use_vec_loss True \
  -att_func 'dot'

sim's People

Contributors

Stargazers

Watchers

sim's Issues

DIEN not available for the SIM model

Hi,

We are trying to reproduce the paper's accuracy fir SIM soft on Amazon dataset. In the paper there is AUC ROC = 0.7510, however when running the code in this repo we get only 0.7221. This is probably we use DIN whereas in the paper there is DIEN used in second stage of the model. Will you provide DIEN model in this repo to fully reproduce the original's paper accuracy?

tttwwy / sim Goto Github PK