FICLRec: Frequency enhanced intent contrastive learning for sequential recommendation

User purchasing behavior is mainly driven by their intentions. However, existing methods typically favor low-frequency intents, leading to insufficient capability in capturing more expressive high-frequency intents. Moreover, like typical sequence recommendations, data sparsity remains a primary factor influencing recommendation performance. To address this issue, we propose a Frequency Enhanced Intent Contrastive Learning Recommendation model (FICLRec), which innovatively utilizes frequency information from users' latent intentions to improve the recognition of high-frequency intents. Additionally, we introduce frequency contrastive learning to reduce the negative impact of data sparsity on model performance. To validate the effectiveness of the proposed method, extensive experiments were conducted on five real-world datasets: Beauty (0.19M interactions), Sports (0.29M interactions), Toys (0.16M interactions), Yelp (0.31M interactions), and LastFM (0.05M interactions). The experimental results indicate that, in comparison with baseline models, our method improves by 2.03%, 4.87%, 2.50%, 13.85%, and 16.93% on five datasets, proving the effectiveness of our method. Our implemented model is available via https://github.com/syf1844803351/FICLRec.