With the advancement of the Internet, digital assaults are changing quickly and the network safety circumstance isn't idealistic. AI (ML) and Deep Learning (DL) techniques for network investigation of interruption identification and gives a short instructional exercise portrayal of every ML/DL strategy. Papers addressing every technique were listed, perused, and summed up dependent on their fleeting or warm connections. Since information are so significant in ML/DL strategies, they portray a portion of the regularly utilized organization datasets utilized in ML/DL, examine the difficulties of utilizing ML/DL for network protection and give recommendations to explore headings. The KDD informational collection is a notable benchmark in the exploration of Intrusion Detection strategies. A ton of work is continuing for the improvement of interruption recognition procedures while the exploration on the information utilized for preparing and testing the location model is similarly of prime concern on the grounds that better information quality can improve disconnected interruption discovery. This venture presents the investigation of KDD informational collection concerning four classes which are Basic, Content, Traffic and Host in which all information ascribes can be arranged utilizing MODIFIED RANDOM FOREST(MRF). The investigation is finished regarding two unmistakable assessment measurements, Detection Rate (DR) and False Alarm Rate (FAR) for an Intrusion Detection System (IDS). Because of this exact investigation on the informational index, the commitment of every one of four classes of qualities on DR and FAR is demonstrated which can help improve the appropriateness of informational index to accomplish most extreme DR with least FAR. The exploratory outcomes got indicated the proposed strategy effectively bring 91% arrangement exactness utilizing just 12 chose highlights and 97% order precision utilizing 36 highlights, while each of the 42 preparing highlights accomplished 98% grouping precision.