Border Gateway Protocol (BGP) anomalies have interrupted network connection on a large scale, henceforth; recognizing them is of very important. Machine learning algorithms play a vital role in detecting the BGP anomalies in network. This paper intends to propose a new BGP anomaly detection method under certain processes (i) Feature Extraction and (ii) Classification. In feature extraction, certain features like "Number of Exterior Gateway Protocol (EGP) packets, Number of Interior Gateway Protocol (IGP) packets, Number of incomplete packets, Maximum Autonomous System (AS) path length, average AS-path length, packet size" etc. are extracted. Along with this, the statistical features such as mean, mode, variance, median, standard deviation, and higher-order statistical features such as kurtosis, skewness, second moment, entropy, and percentiles are also extracted. Subsequently, the classification is carried out by a hybrid classifier model that merges the Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) models.