RNN / LSTM with modified Adam optimizer in deep learning approach for automobile spare parts demand forecasting

The spare parts demand forecasting is very much essential for the organizations to minimize the cost and prevent the stock outs. The demand of spare parts/ car sales distribution is an important factor in inventory control. The valuation of the demand is challenging as the automobile spare parts/car sales demand are often recurrent. The renowned empirical method adopts historical demand data to create the distribution of lead time demand. Although it works reasonably well when service requirements are relatively low, it has difficulty reaching high target service levels. In this paper, we proposed Recurrent Neural Networks/ Long-Short Term Memory (RNN / LSTM) with modified Adam optimizer to predict the demand for spare parts. In this LSTM, weight vectors are generated respectively. These weights are optimized using the Modified-Adam algorithm. The accuracy of the forecast and the performance of the inventory are considered in the experimental result. Experimental results confirm that RNN / LSTM with a Modified-Adam works well with minimal error compared to other existing methods. We conclude that the proposed RNN/LSTM with Modified-Adam algorithm is well suited for the prediction of automobile spare parts.

     

Maximizing a deep submodular function optimization with a weighted MAX-SAT problem for trajectory clustering and motion segmentation

Computer vision models are commonly defined for maximum constrained submodular functions lies at the core of low-level and high-level models. In such that, the pixels that are to be grouped or segmenting moving object remains a challenging task. This paper proposes a joint framework for maximizing submodular energy subject to a matroid constraint using Deep Submodular Function (DSF) optimization approximately to solve the weighted MAX-SAT (Maximum Satisfiability) problem and a new trajectory clustering method called Simple Slice Linear clustering (SSLIC) and motion cue method for trajectory clustering and motion segmentation. In this objective function, the illustrative trajectories of a small number are selected automatically by deep submodular maximization. Although, the exploitation of monotone and submodular properties are further maximized and the complexity is reduced by a continuous greedy algorithm. The bound guarantees a fully sliced curve of (1- S/e) to (1–1/e) with less running time. Lastly, the motion is segmented by the motion cue method to accurately differentiate the set of frames for different scenes. Experiments on the Hopkins 155, Berkley Motion Segmentation (BMS) and FBMS-59 datasets display the trajectory clustering and motion segmentation result over its superior performance with respect to 14 quality evaluation metrics. Hence the simulation result shows that the proposed joint framework attains better performance than existing methods on trajectory clustering and motion segmentation task.