Customer reviews in online websites has been increased a lot nowadays. Detecting aspects on those reviews are becoming a challenging task because of size complexity. Hence, an automated mechanism is needed to detect the product aspects from the online consumer reviews. In this paper we modeled an unsupervised technique to detect product aspects. In general, the product aspect may be single word or multiple words. To detect single word aspects term dependency analysis is done in which aspects are extracted based on their opinions. Multiword aspects are determined by using a technique called Frequency and Length based aspect relation (FLAT) which is extended from the technique of c-value. In order to increase the precision and recall, two pruning strategies are followed. Finally, an enhanced bootstrapping affinity measure algorithm, Frequency and Inter-relation score (FIR) has been formulated to determine the co-occurrence between the aspects and the seed aspects. Evaluation of the proposed model is carried out for electronic products by using the product reviews provided on online review websites.