DATA ANALYSIS OF CONSUMER COMPLAINTS IN BANKING INDUSTRY USING K-MEANS CLUSTERING AND SENTIMENTAL ANALYSIS

Abstract : This paper specialise in exploring and analyzing Consumer Finance Complaints data, to seek out what percentage similar complaints are there in reference to an equivalent bank or service or product. These datasets fall into the complaints of Credit reporting, Mortgage, Debt Collection, personal loan and Banking Accounting. By using data processing techniques, cluster analysis also as predictive modeling is applied to get valuable information about complaints in certain regions of the Country. The banks that are receiving customer complaints filed against them will analyse the complaint data to supply results on where the foremost complaints are being filed, what products/ services are producing the foremost complaints and other useful data. Our model will assist banks in identifying the situation and kinds of errors for resolution, resulting in increased customer satisfaction to drive revenue and profitability.

 EXISTING SYSTEM :

 ? The majority of existing SA models analyze textual data gathered from social media while ignoring direct client/company interactions. ? Such transformations are not only helping the service providers but also increasing the risk of maintaining the existing customers from the attractive offers given by the competitors. ? This model takes much less computation time for performing the prediction operation on the customer churn as compared to the existing models.

 DISADVANTAGE :

 ? A consumer’s complaints present bank or reporting agency with an opportunity to identify and rectify specific problems with their current product or service. ? Clustering is regarded as a crucial unsupervised learning problem, that tries to search for similar structures among an unlabeled data set. ? This is an interesting problem as well as the evaluation of incremental learning algorithms themselves for which, as underlined in, there are still no shared solutions, especially for non-stationary data.

 PROPOSED SYSTEM :

 • We present a contextual investigation of stand-type forced air systems to outline how the proposed strategy will characterize client necessities. • In, a framework for email sentiment analysis is proposed, based on a hybrid scheme combining k-means clustering and an SVM classifier. • The proposed solution is particularly useful for gradually improving the initial model while avoiding the costs of training a new model from scratch, without deteriorating the prediction accuracy over time.

 ADVANTAGE :

 ? Reflecting on our cluster modeling results, we evaluated the performance of our model. ? Through the proposed approach, in fact, it is possible to start using the pre-trained system to operate immediately (without initial training) and, therefore, to perform a gradual improvement by exploiting the operator’s feedback until optimal performance is achieved. ? Indeed, cross-validation gives the model the opportunity to train on multiple training-test splits providing a better indication of the model’s performance on unseen data.