Interplanetary scintillation (IPS) is a radio scattering phenomenon generated by the disturbances in the solar wind. Nagoya University conducts IPS observations in the 327-MHz band using three radio telescopes in Japan and calculates the solar wind speed by analyzing the correlation of the observed signals. In recent years, growing artificial radio sources have intensified radio frequency interference (RFI), degrading data quality and observation efficiency. Traditional threshold-based RFI detection methods are no longer sufficient for handling the diverse and complex characteristics of modern RFI. The first objective of this study is developing a model to detect pulse-type RFI using one-dimensional time-series intensity data obtained via square-law detection. This flagging system requires an accuracy of at least 95.80% for signals whose intensities exceed three times the standard deviation. RFI-free IPS signals are defined based on the model proposed by Mejia-Ambriz et al. (2015) and observational parameters from Nagoya University. The solar wind velocity distributions of the RFI-free IPS data were consistent with typical solar wind velocity distributions. As a first step, we used an unsupervised Long Short-Term Memory Autoencoder (LSTM AE) for pulse-type RFI detection. In our design, the model preserves temporal structure in the latent space rather than compressing signals into a single vector. While it achieved a median F1 score of 0.97 for identifying time intervals containing RFI, its ability to localize individual anomalous points was limited, with a median F1 score below 0.10. Next, supervised learning models—including a Convolutional Neural Network (CNN), U-Net, Temporal Convolutional Network (TCN), and Transformer—consistently achieved F1 scores above 0.99, with the CNN performing best. These results demonstrate that supervised approaches offer substantial advantages in detecting complex pulse-type RFI. As future work, we will focus on detecting burst-type RFI and developing models capable of separating RFI from observed IPS signals.