A hybrid method for forecasting new product sales based on fuzzy clustering and deep learning

Purpose The purpose of this paper is to solve the problem of low accuracy in new product demand forecasting caused by the absence of historical data and inadequate consideration of influencing factors. Design/methodology/approach A hybrid new product demand forecasting model combining clustering analysis and deep learning is proposed. Based on the product similarity measurement, the weight of product similarity attributes is realized by using the method of fuzzy clustering-rough set, which provides a basis for the acquisition and collation of historical sales data of similar products and the determination of product similarity. Then the prediction error of Bass model is adjusted based on similarity through a long short-term memory neural network model, where the influencing factors such as product differentiation, seasonality and sales time on demand forecasting are embedded. An empirical example is given to verify the validity and feasibility of the model. Findings The results emphasize the importance of considering short-term impacts when forecasting new product demand. The authors show that useful information can be mined from similar products in demand forecasting, where the seasonality, product selling cycles and sales dependencies have significant impacts on the new product demand. In addition, they find that even in the peak season of demand, if the selling period has nearly passed the growth cycle, the Bass model may overestimate the product demand, which may mislead the operational decisions if it is ignored. Originality/value This study is valuable for showing that with the incorporation of the evaluation method on product similarity, the forecasting model proposed in this paper achieves a higher accuracy in forecasting new product sales.

Download Full-text

Consumers’ pre-launch awareness and preference on movie sales

European Journal of Marketing ◽

10.1108/ejm-06-2015-0324 ◽

2016 ◽

Vol 50 (5/6) ◽

pp. 1024-1046 ◽

Cited By ~ 6

Author(s):

Sangkil Moon ◽

Junhee Kim ◽

Barry L. Bayus ◽

Youjae Yi

Keyword(s):

New Product ◽

Forecasting Model ◽

Content Type ◽

Box Office ◽

Consumption Frequency ◽

Marketing Resources ◽

Low Performance ◽

Product Sales ◽

Negative Sentiment ◽

Product Preference

Purpose The purpose of this paper is to provide insightful advice that can improve the practice of using consumers’ pre-launch awareness and preference (AP) changes to predict the sales of new movies. Design/methodology/approach This paper applies a new movie box-office revenue forecasting model based on consumers’ weekly AP measures, to take advantage of the industry’s practice of using weekly survey data containing the AP measures of upcoming new movies. Specifically, a sales forecasting model is developed on the basis of the theory that the combination of the nature of new product preference (positive vs negative) and the timing of new product awareness (early vs recent) influences entertainment product sales. Findings This paper shows that early awareness consumers are as important as late awareness consumers in determining new product sales, suggesting that more marketing resources need to be allocated earlier than currently practiced. This paper also shows that when negative preferences dominate positive preferences well ahead of a product’s release, marketing efforts cannot overcome the negative sentiment of the market. Finally, the empirical application illustrates that three consumer segments varying in product expertise and consumption frequency reveal different AP patterns among high-, medium- and low-performance products. Originality/value This paper is intended to provide insightful advice that can improve the AP-based approach in entertainment industries. Toward that end, the authors emphasize two major aspects in association with new entertainment product sales: rethinking survey-based AP measures and examining heterogeneous consumer segments’ differential AP patterns.

Download Full-text

Pre-launch new product demand forecasting using the Bass model: A statistical and machine learning-based approach

Technological Forecasting and Social Change ◽

10.1016/j.techfore.2013.08.020 ◽

2014 ◽

Vol 86 ◽

pp. 49-64 ◽

Cited By ~ 34

Author(s):

Hakyeon Lee ◽

Sang Gook Kim ◽

Hyun-woo Park ◽

Pilsung Kang

Keyword(s):

Machine Learning ◽

Demand Forecasting ◽

New Product ◽

Bass Model ◽

Product Demand

Download Full-text

Usage of the Pareto Fronts as a Tool to Select Data in the Forecasting Process—A Short-Term Electric Energy Demand Forecasting Case

Energies ◽

10.3390/en14113204 ◽

2021 ◽

Vol 14 (11) ◽

pp. 3204

Author(s):

Michał Sabat ◽

Dariusz Baczyński

Keyword(s):

Energy Demand ◽

Demand Forecasting ◽

Electric Energy ◽

Forecasting Model ◽

Short Term ◽

Complex Evaluation ◽

Proposed Model ◽

Energy Demand Forecasting ◽

Pareto Fronts ◽

Naive Approach

Transmission, distribution, and micro-grid system operators are struggling with the increasing number of renewables and the changing nature of energy demand. This necessitates the use of prognostic methods based on ever shorter time series. This study depicted an attempt to develop an appropriate method by introducing a novel forecasting model based on the idea to use the Pareto fronts as a tool to select data in the forecasting process. The proposed model was implemented to forecast short-term electric energy demand in Poland using historical hourly demand values from Polish TSO. The study rather intended on implementing the range of different approaches—scenarios of Pareto fronts usage than on a complex evaluation of the obtained results. However, performance of proposed models was compared with a few benchmark forecasting models, including naïve approach, SARIMAX, kNN, and regression. For two scenarios, it has outperformed all other models by minimum 7.7%.

Download Full-text

An Ensemble Energy Consumption Forecasting Model Based on Spatial-Temporal Clustering Analysis in Residential Buildings

Energies ◽

10.3390/en14113020 ◽

2021 ◽

Vol 14 (11) ◽

pp. 3020

Author(s):

Anam-Nawaz Khan ◽

Naeem Iqbal ◽

Atif Rizwan ◽

Rashid Ahmad ◽

Do-Hyeun Kim

Keyword(s):

Cluster Analysis ◽

Deep Learning ◽

Residential Buildings ◽

Ensemble Forecasting ◽

Forecasting Model ◽

Ensemble Model ◽

Learning Models ◽

Short Term ◽

Floor Level ◽

Consumption Data

Due to the availability of smart metering infrastructure, high-resolution electric consumption data is readily available to study the dynamics of residential electric consumption at finely resolved spatial and temporal scales. Analyzing the electric consumption data enables the policymakers and building owners to understand consumer’s demand-consumption behaviors. Furthermore, analysis and accurate forecasting of electric consumption are substantial for consumer involvement in time-of-use tariffs, critical peak pricing, and consumer-specific demand response initiatives. Alongside its vast economic and sustainability implications, such as energy wastage and decarbonization of the energy sector, accurate consumption forecasting facilitates power system planning and stable grid operations. Energy consumption forecasting is an active research area; despite the abundance of devised models, electric consumption forecasting in residential buildings remains challenging due to high occupant energy use behavior variability. Hence the search for an appropriate model for accurate electric consumption forecasting is ever continuing. To this aim, this paper presents a spatial and temporal ensemble forecasting model for short-term electric consumption forecasting. The proposed work involves exploring electric consumption profiles at the apartment level through cluster analysis based on the k-means algorithm. The ensemble forecasting model consists of two deep learning models; Long Short-Term Memory Unit (LSTM) and Gated Recurrent Unit (GRU). First, the apartment-level historical electric consumption data is clustered. Later the clusters are aggregated based on consumption profiles of consumers. At the building and floor level, the ensemble models are trained using aggregated electric consumption data. The proposed ensemble model forecasts the electric consumption at three spatial scales apartment, building, and floor level for hourly, daily, and weekly forecasting horizon. Furthermore, the impact of spatial-temporal granularity and cluster analysis on the prediction accuracy is analyzed. The dataset used in this study comprises high-resolution electric consumption data acquired through smart meters recorded on an hourly basis over the period of one year. The consumption data belongs to four multifamily residential buildings situated in an urban area of South Korea. To prove the effectiveness of our proposed forecasting model, we compared our model with widely known machine learning models and deep learning variants. The results achieved by our proposed ensemble scheme verify that model has learned the sequential behavior of electric consumption by producing superior performance with the lowest MAPE of 4.182 and 4.54 at building and floor level prediction, respectively. The experimental findings suggest that the model has efficiently captured the dynamic electric consumption characteristics to exploit ensemble model diversities and achieved lower forecasting error. The proposed ensemble forecasting scheme is well suited for predictive modeling and short-term load forecasting.

Download Full-text

High frequency short-term demand forecasting model for distribution power grid based on ARIMA

2012 IEEE International Conference on Computer Science and Automation Engineering (CSAE) ◽

10.1109/csae.2012.6272958 ◽

2012 ◽

Cited By ~ 4

Author(s):

Hongming He ◽

Tao Liu ◽

Ruimin Chen ◽

Yong Xiao ◽

Jinfeng Yang

Keyword(s):

High Frequency ◽

Power Grid ◽

Demand Forecasting ◽

Forecasting Model ◽

Short Term ◽

Demand Forecasting Model ◽

Grid Based

Download Full-text

Day-ahead load forecasting using improved grey Verhulst model

Journal of Engineering Design and Technology ◽

10.1108/jedt-12-2019-0337 ◽

2020 ◽

Vol 18 (5) ◽

pp. 1335-1348

Author(s):

Ariel Mutegi Mbae ◽

Nnamdi I. Nwulu

Keyword(s):

Power Systems ◽

Unit Commitment ◽

Load Forecasting ◽

Forecasting Model ◽

Short Term ◽

Content Type ◽

Dispatch System ◽

System Maintenance ◽

Electricity Load ◽

Generator Unit

Purpose In the daily energy dispatch process in a power system, accurate short-term electricity load forecasting is a very important tool used by spot market players. It is a critical requirement for optimal generator unit commitment, economic dispatch, system security and stability assessment, contingency and ancillary services management, reserve setting, demand side management, system maintenance and financial planning in power systems. The purpose of this study is to present an improved grey Verhulst electricity load forecasting model. Design/methodology/approach To test the effectiveness of the proposed model for short-term load forecast, studies made use of Kenya’s load demand data for the period from January 2014 to June 2019. Findings The convectional grey Verhulst forecasting model yielded a mean absolute percentage error of 7.82 per cent, whereas the improved model yielded much better results with an error of 2.96 per cent. Practical implications In the daily energy dispatch process in a power system, accurate short-term load forecasting is a very important tool used by spot market players. It is a critical ingredient for optimal generator unit commitment, economic dispatch, system security and stability assessment, contingency and ancillary services management, reserve setting, demand side management, system maintenance and financial planning in power systems. The fact that the model uses actual Kenya’s utility data confirms its usefulness in the practical world for both economic planning and policy matters. Social implications In terms of generation and transmission investments, proper load forecasting will enable utilities to make economically viable decisions. It forms a critical cog of the strategic plans for power utilities and other market players to avoid a situation of heavy stranded investment that adversely impact the final electricity prices and the other extreme scenario of expensive power shortages. Originality/value This research combined the use of natural logarithm and the exponential weighted moving average to improve the forecast accuracy of the grey Verhulst forecasting model.

Download Full-text

Ultra-Short-Term Load Demand Forecast Model Framework Based on Deep Learning

Energies ◽

10.3390/en13184900 ◽

2020 ◽

Vol 13 (18) ◽

pp. 4900 ◽

Cited By ~ 1

Author(s):

Hongze Li ◽

Hongyu Liu ◽

Hongyan Ji ◽

Shiying Zhang ◽

Pengfei Li

Keyword(s):

Time Series ◽

Deep Learning ◽

Meteorological Data ◽

Demand Forecasting ◽

Dense Layer ◽

Load Prediction ◽

Demand Forecast ◽

Short Term ◽

Model Framework ◽

Load Demand

Ultra-short-term load demand forecasting is significant to the rapid response and real-time dispatching of the power demand side. Considering too many random factors that affect the load, this paper combines convolution, long short-term memory (LSTM), and gated recurrent unit (GRU) algorithms to propose an ultra-short-term load forecasting model based on deep learning. Firstly, more than 100,000 pieces of historical load and meteorological data from Beijing in the three years from 2016 to 2018 were collected, and the meteorological data were divided into 18 types considering the actual meteorological characteristics of Beijing. Secondly, after the standardized processing of the time-series samples, the convolution filter was used to extract the features of the high-order samples to reduce the number of training parameters. On this basis, the LSTM layer and GRU layer were used for modeling based on time series. A dropout layer was introduced after each layer to reduce the risk of overfitting. Finally, load prediction results were output as a dense layer. In the model training process, the mean square error (MSE) was used as the objective optimization function to train the deep learning model and find the optimal super parameter. In addition, based on the average training time, training error, and prediction error, this paper verifies the effectiveness and practicability of the load prediction model proposed under the deep learning structure in this paper by comparing it with four other models including GRU, LSTM, Conv-GRU, and Conv-LSTM.

Download Full-text

A demand forecasting model based on the improved Bass model for fast fashion clothing

International Journal of Clothing Science and Technology ◽

10.1108/ijcst-08-2019-0114 ◽

2020 ◽

Vol ahead-of-print (ahead-of-print) ◽

Author(s):

Xiaoxi Zhou ◽

Jianfei Meng ◽

Guosheng Wang ◽

Qin Xiaoxuan

Keyword(s):

Historical Data ◽

Consumer Preferences ◽

Demand Forecasting ◽

Impact Factors ◽

Bass Model ◽

Demand Forecast ◽

Fast Fashion ◽

Content Type ◽

Sales Data ◽

Demand Forecasting Model

PurposeThis paper examines the problem of lack of historical data and inadequate consideration of factors influencing demand in the forecasting of demand for fast fashion clothing and proposes an improved Bass model for the forecasting of such a demand and the demand for new clothing products.Design/methodology/approachFrom the perspective of how to solve the lack of data and improve the precision of the clothing demand forecast, this paper studies the measurement of clothing similarity and the addition of demand impact factors. Using the fuzzy clustering–rough set method, the degree of resemblance of clothing is determined, which provides a basis for the scientific utilisation of historical data of similar clothing to forecast the demand for new clothing. Besides, combining the influence of consumer preferences and seasonality on demand forecasting, an improved Bass model for a fast fashion clothing demand forecast is proposed. Finally, with a forecasting example of demand for clothing, this study also tests the validity of the method.FindingsThe objective measurement method of clothing similarity in this paper solves the problem of the difficult forecasting of demand for fast fashion clothing due to a lack of sales data at the preliminary stage of the clothing launch. The improved Bass model combines, comprehensively, consumer preferences and seasonality and enhances the forecast precision of demand for fast fashion clothing.Originality/valueThe paper puts forward a scientific, quantitative method for the forecasting of new clothing products using historical sales data of similar clothing, thus solving the problem of lack of sales data of the fashion.

Download Full-text