jailukanna/cdap-b

PROJECT TITLE - Supermarket retail-based demand and price prediction of vegetables

SUPERVISOR - Dr. Dharshana Kasthurirathna

CO-SUPERVISOR - Ms. Thilini Jayalath

MAIN OBJECTIVE - This system will give time to farmer to grow the needed product when the stock level starting to get lower. Price and the demand prediction minimise the wastage of vegetables. Because farmers grow vegetables according to the demand. So that when the stock is in the lower level, farmer can provide the needed product without any inconvenience. Therefore, waste will be minimum, and warehouse won’t face any product shortage. Stock level alert will grant using machine learning. When releasing a stock to supermarket execute a function and get the alert through machine learning model. Then after making a prediction using another model to get the stock count which wants for next session. Those predicted counts will send to the selected farmers. Farmers will select by using their ratings. (Using an optimization algorithm). If this solution is used to predict the retail price, farmers income can be increased since they can manage their cultivation according to the predictions. When the demand is predicted, the wastage inside a warehouse will be minimum. So that the income and the profit of the company will be increased. Since the warehouse stock keeping is automated, a stock keeper won’t be needed anymore. When logging the stock details using an automated system, errors can be minimized so that the productivity of the company goes high.

MAIN RESEARCH QUESTIONS - When a stock level is getting decreased inside a warehouse, this system will predict it before it gets lower. Prediction is done through considering the vegetable growth period. When the system detects about a decreasing stock level, it will consider the time it takes to grow that vegetable and then notify the supplier, typically the farmer. The relevant farmer will be selected by considering the best farmer through a rating method. This system will give time to farmer to grow the needed product when the stock level starting to get lower. So that when the stock is in the lower level, farmer can provide the needed product without any inconvenience. Therefore, waste will be minimum. Our research solution can be used to predict the retail price, so that farmers income can be increased since they can manage their cultivation according to the predictions. When the demand is predicted, the wastage inside a warehouse will be minimum. So that the income and the profit of the company will be increased. Since the warehouse stock keeping is automated, a stock keeper won’t be needed anymore. When stock logging is automated, errors can be minimized, and the productivity of a company goes high. In certain period of times in Sri Lanka, vegetable wastage can be identified. It can be either a wastage or a shortage of the certain vegetable. Farmers may over produce a certain vegetable without knowing the actual amount which its needed for that time period. So, when they over produced that using all their hard-earned money, and time eventually it will become a wastage. On the other hand, the buyer who is buying that vegetable will be in a big trouble also. Which they will buy that vegetable without knowing the actual amount needed. That's a wastage. This scenario is visibly shown inside a supermarket.

INDIVIDUAL RESERACH QUESTION -

IT17136402 - In certain period of times in Sri Lanka, vegetable wastage can be identified. It can be either a wastage or a shortage of certain vegetable. Farmers may over produce a certain vegetable without knowing the actual amount which isneeded for that time period. So, when they over produced that using all their hard-earned money, and time eventually it will become a wastage. When farmers can't sell
their harvest for a reasonable price, either they let harvest to be rot in the fields or in the worst-casescenario,they give up on their lives even. On the other hand, the buyer who is buying that vegetable will be in a big trouble also. Which they will buy that vegetable without knowing the actual amount needed. This wastage is occurring mainly because of the price fluctuation in perishable crops like tomato, chili etc.

IT17049382 - In supply chain management, accurate prediction of goods demand is an important element because based on predicted demand we can optimize the level of stock. The demand for goods may differ affected by numerous factors including weather, seasonal effects and price elasticity. In recent years the distribution of fresh vegetables has become an important concern matter.From farmer’s perspective inability to predict the demand for vegetables stands as the main reason for vegetable wastage. Both farmers and the supermarket face lots of inconvenience due to the lack of prediction of vegetable demand. When considering the retailer’s perspective, due to unpredicted demand, time to time the warehouse can be in overstocked and understocked conditions. That put a major impact on the company revenue.So it is needed to have a predicted model to reduce the wastage of vegetables.Here especially we have to consider the fact weather.Because according to the different weather conditions the quantity demand may vary.When considering the factor weather regional difference plays a major role, as different areas of the country have different types of weather conditions.Due to that reason in the different areas of the country, there are some variations in growing the vegetables.We should understand the seasonal effects and weather as they are the highly affecting areas in demand prediction.Another impact is import foods, promotions and company productions of the vegetables.So that the quantity demand of a certain vegetable may vary accordingly.Local demand is also another considering fact in this problem. In supermarkets, farmers should be informed about the needed quantity of a certain vegetable at least two days before because usually, they do the prediction daily.So toorder the exact quantity of the exact vegetable item, we must have a way to predict the needed quantity precisely.So that we can reduce the wastage of vegetables. Another impact is the availability and price. They are highly considered factors in demand prediction. It is a must to consider the quantity that farmers can supply. It can be shown as another impact of demand prediction. This system will give time to the farmer to grow the needed product when the stock level starting to get lower.So that whenthe stock level is at the lower level, the farmer can provide the needed products without any inconvenience. Therefore the wastage of the vegetables will reduce.Because when the demand is predicted the wastage in the warehouse becomes minimum. Since the warehouse stock keeping is automated, there won’t be needed for a stock keeper.

IT17003438 - With this approach, farmers are directly associated with the supermarket warehouse. Without best satisfactory farmers, unbearable to produce high-quality, low-price products. When a stock level is getting decreased inside a warehouse, this system will predict it before it gets lower. Prediction is done by considering the vegetable growth period. When the system detects about a decreasing stock level, it will consider the time it takes to grow that vegetable and then notify the supplier, typically the farmer.The relevant farmer will be selected by considering the best farmer through a rating method. For that, to be the best farmer, they mustbe able to better understand what the best product is. Farmers also need to knowat that time, what the price and demand of the products are. And managers also need to know the quality of the products. So that when the stock is at the lower level, farmers can provide the needed product without any inconvenience. Therefore, waste will be minimized. Farmers may overproduce a certain vegetable without knowing the actual amount which it needed for that time period. So, when they overproduced that using all their hard-earned money, and time eventually it will become a wastage. On the other hand, the buyer who is buying that vegetable will be in big trouble also. Which they will buy that vegetable without knowing the actual amount needed. That's a wastage. This scenario is visibly shown inside a supermarket.

IT17016162 - Nowadays the consumers are highly interested in the origin of the vegetables they purchase. Whenever a consumer is purchasing vegetables or any sort of food, they tend to question whether the vegetables are organic as they claim to be by the supermarket. Therefore with this research component, we address the requirement of verifying the source of vegetables by applying blockchain technology in the supermarket supply chain.

INDIVIDUAL OBJECTIVES -

IT17136402 - This research will predict prices of some crash crops using Time Series Prediction through Regression Approach. This system will give time to farmer to grow the needed product when the stock level starting to get lower. So that when the stock is in the lower level,
farmer can provide the needed product without any inconvenience. Therefore, waste will be minimum. This research solution can be used to predict the retail price, so that farmers income can be increased since they can manage their cultivation
according to the predictions.Also,it helps the Government to take effective decisions like planning agricultural
development programs, take decisions on exporting and importing crops. This solution can be used to minimize
the wastage inside a warehouse. So that the income and the profit of the company will be increased. Since stock logging is automated, a stock-keeper won't be needed, errors can be minimized, productivity and the revenue of a company goes high.

IT17049382 - By analyzing daily sales data of consumer goods collected from Point Of Sale System (POS) of Cargills (Ceylon) PLC. Here we select Cargills (Ceylon) PLC. because they comprise records of price and quantity of each item extend over several years. In the opinion of regional preferences and seasonal effects, demand may change accordingly. We proposed to predict a model by estimating the demand curve of vegetables by employing a regression approach and neural networks. We aim to show that there are regional differences and seasonal effects especially effects of weather effects to differ the demand level. Stock level alert will grant using machine learning. When releasing a stock to supermarket execute a function and get the alert through the machine learning model. Then after making a prediction using another model to get the stock count which wants for the next session. Those predicted counts will send to the selected farmers. So the farmers will grow vegetables according to the demand. When the demand is predicted, the wastage inside the warehouse will be minimum.

IT17003438 - The objective of this is how managers actually select the best farmers. According to farmer’s performance, best farmers will select by using their ratings. This solution is used to predict the retail price. If this solution is used to predict retail price, farmers’income can be increased since they can manage their cultivation according to the prediction. And, there are many wastages in the warehouse, because of not correct management in the cultivation. When the demand is predicted using this solution, will be minimized wastage of goods inside the warehouse. And therefore,the income and profit if the company will be increased.

IT17016162 - This research component allows the consumers to tack and view the history up to the origin of a selected vegetable in order to verify the vegetable quality and to verify whether the vegetable is suitable for consumption by simply scanning a QR code which is provided for the vegetables.For the above mentioned purpose Hyperledger Sawtoothis proposed to be used as the blockchain,
which is a private blockchain that is issued with writing permissions and rules to guarantee access only for members that are recognized as legitimate participants in the process.