A method and system for evaluating enterprise annual R&D expense budget based on a web platform

Abstract

The present invention proposes a method and system for evaluating an enterprise's annual R&D expense budget based on a web page platform, wherein the method comprises: building a user interface through Vue.js and creating a data entry form to obtain the user's historical R&D expense segmentation data and verify the format of the historical R&D expense segmentation data; obtaining preliminary predicted R&D expense data based on the historical R&D expense segmentation data; predicting the future development trend of the project based on project keywords, and obtaining a total evaluation budget based on the preliminary prediction and future development trend, thereby providing a scientific, comprehensive and efficient R&D budget evaluation tool for enterprises.

Description

Technical Field

[0001] The present invention proposes a method and system for evaluating an enterprise's annual R&D expense budget based on a webpage platform, and relates to the technical field of expense evaluation. Background Art

[0002] With the rapid development of science and technology and increasingly fierce market competition, companies are paying more and more attention to R&D investment. Reasonable and scientific evaluation and allocation of annual R&D expenses can effectively enhance a company's innovation capabilities and market competitiveness. However, traditional R&D expense budget evaluation methods often rely on manual judgment and simple historical data analysis, which makes it difficult to fully and objectively reflect the potential and development trends of the project. Data is scattered and non-standardized: The company's historical R&D expense data is scattered in different systems and files, lacking a unified format and specifications, resulting in time-consuming and labor-intensive data collection and organization. Lack of scientific forecasting methods: Traditional evaluation methods are mostly based on experience and simple historical data analysis, which makes it difficult to accurately predict future market and technological development trends. Weak dynamic adjustment capabilities: The market environment and technological trends change rapidly, and traditional static budget evaluation methods cannot respond in a timely manner, making it difficult to make dynamic adjustments. Summary of the Invention

[0003] The present invention provides a web-based platform-based enterprise annual R&D expense budget assessment method and system to solve the above-mentioned problems:

[0004] The present invention proposes a method for evaluating an enterprise's annual R&D expense budget based on a webpage platform, the method comprising:

[0005] Build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data;

[0006] Obtain preliminary forecast R&D expense data based on the historical R&D expense breakdown data;

[0007] Predict the future development trend of the project based on project keywords, and obtain the total evaluation budget based on the preliminary prediction and future development trend.

[0008] Furthermore, a user interface is built using Vue.js, and a data entry form is created to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data, including:

[0009] Create data entry forms with HTML and CSS, and build user interfaces using Vue.js in JavaScript;

[0010] Use the JavaScript library Yup on the front end to perform preliminary verification to ensure that the data entered by the user meets the requirements;

[0011] Submit the form data to the backend through the Fetch API;

[0012] The backend service receives the data and further verifies and processes the historical R&D expense breakdown data rows.

[0013] Furthermore, preliminary forecast R&D expense data is obtained based on the historical R&D expense breakdown data, including:

[0014] Obtain R&D expense data for the past five years and address missing data and outliers in the R&D expense data;

[0015] Using the data from the past 5 years, we train an ARIMA model and select the best p and q values. The ARIMA model is:

[0016] Y t =c+φ1Y t-1 +φ2 Y t-2 +φ p Y t-p +…+θ1ε t-1 +…+θ q ε t-q +ε t

[0017] Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation;

[0018] Use the trained ARIMA model to predict R&D expenses;

[0019] Based on the output of the model, the prediction results are displayed through the Node.js visualization interface.

[0020] Furthermore, the optimal p and q values are selected, including:

[0021] Draw a line graph of the historical R&D expense data obtained;

[0022] The stationarity of the time series is tested by the ADF test method, and the non-stationary time series is subjected to difference processing;

[0023] Draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) to preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part.

[0024] Use the AIC and BIC indicators to select the optimal combination of p and q, traverse different p and q values, compare the AIC and BIC values of each model, and select the combination that minimizes AIC or BIC to determine the optimal p and q values.

[0025] Furthermore, the future development trend of the project is predicted based on the project keywords. Based on the preliminary prediction and future development trend, the total evaluation budget is obtained, including:

[0026] Obtain the evaluation project categories, design a web crawler using Python's Scrapy library, and crawl articles and post data on the forum;

[0027] Organize the collected data by time and count the number of articles and posts in each month;

[0028] Categorize article and post content to identify content relevant to the assessment project category;

[0029] Calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category;

[0030] Use the weighted moving average method to smooth the growth rate, and then calculate the trend score of the project category based on the weighted moving average growth rate;

[0031] The total annual R&D expenditure of the assessed enterprise is obtained based on the trend score and the project expenditure predicted by ARIMA.

[0032] The present invention proposes a web-based platform-based enterprise annual R&D expense budget assessment system, which includes:

[0033] The data acquisition module is used to build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data;

[0034] A preliminary forecasting module, configured to obtain preliminary forecasted R&D expense data based on the historical R&D expense breakdown data;

[0035] The total evaluation module is used to predict the future development trend of the project based on the project keywords, and obtain the total evaluation budget based on the preliminary prediction and future development trend

[0036] Furthermore, the data acquisition module includes:

[0037] Build interface modules for creating data entry forms with HTML and CSS, and building user interfaces using Vue.js in JavaScript;

[0038] The validation format module is used to perform preliminary validation on the front end using the JavaScript library Yup to ensure that the data entered by the user meets the requirements;

[0039] Submit backend module, used to submit form data to the backend through Fetch API;

[0040] The processing module is used for the back-end service to receive data and perform one-step verification and processing on the historical R&D expense breakdown data.

[0041] Furthermore, the preliminary prediction module includes:

[0042] A historical data acquisition module is used to obtain R&D expense data for the past five years and process missing data and outliers in the R&D expense data;

[0043] The training model module is used to train the ARIMA model using the data from the past five years and select the optimal p and q values. The ARIMA model is:

[0044] Y t =c+φ1Y t-1 +φ2 Y t-2 +φ p Y t-p +…+θ1ε t-1 +…+θ q ε t-q +ε t

[0045] Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation;

[0046] The forecasting module is used to forecast R&D expenses using the trained ARIMA model;

[0047] The visualization display module is used to display the prediction results through the Node.js visualization interface based on the output of the model.

[0048] Furthermore, the training model module includes:

[0049] A graph drawing module is used to draw a line graph of the acquired historical R&D expense data;

[0050] The stationary detection module is used to detect the stationarity of the time series through the ADF test method and perform differential processing on non-stationary time series;

[0051] A preliminary determination module is used to draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) graphs, and preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part.

[0052] The module for determining the optimal p and q values is used to select the optimal p and q combination using the AIC and BIC indicators. It traverses different p and q values, compares the AIC and BIC values of each model, and selects the combination that minimizes AIC or BIC to determine the optimal p and q values.

[0053] Furthermore, the overall evaluation module includes:

[0054] The crawling module is used to obtain the evaluation project categories and design a web crawler using Python's Scrapy library to crawl articles and post data on the forum;

[0055] The statistics module is used to organize the collected data by time and count the number of articles and posts in each month;

[0056] Identification module, used to classify the content of articles and posts and identify content related to the assessment item category;

[0057] The growth rate calculation module is used to calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category;

[0058] The trend score calculation module is used to smooth the growth rate using the weighted moving average method, and then calculate the trend score of the project category based on the weighted moving average growth rate;

[0059] The evaluation module is used to obtain the total annual R&D expenditure of the evaluated enterprise based on the trend score and the project expenditure predicted by ARIMA.

[0060] The beneficial effects of the present invention are as follows: resource optimization allocation, through analysis of future trends, can help enterprises allocate resources more efficiently to the most promising projects, thereby increasing the return on R&D investment and the competitiveness of the enterprise; dynamic adaptability, the system design takes into account dynamic input and real-time analysis, can update data and budget forecasts at any time, and help enterprises quickly adapt to market changes; excellent user experience, the front-end framework based on Vue.js can provide a smooth and intuitive operating experience, a friendly user interface and clear data display, reduce the user's operating difficulty, and improve the efficiency and accuracy of data entry. This technical solution can provide enterprises with a scientific, comprehensive and efficient R&D budget assessment tool through the comprehensive use of front-end technology, data analysis and prediction models. Through accurate data entry verification, trend forecasting and budget calculation, enterprises can optimize R&D resource allocation based on actual data and market trends, better plan future investments, enhance market competitiveness and promote long-term development. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a schematic diagram of a web-based platform-based enterprise annual R&D expense budget assessment method according to the present invention. DETAILED DESCRIPTION

[0062] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein may be combined with each other.

[0063] The following description sets forth numerous specific details to facilitate a thorough understanding of the present invention. The embodiments described are merely a portion of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0065] One embodiment of the present invention provides a method for evaluating an enterprise's annual R&D expense budget based on a webpage platform, the method comprising:

[0066] Build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data;

[0067] Obtain preliminary forecast R&D expense data based on the historical R&D expense breakdown data;

[0068] Predict the future development trend of the project based on project keywords, and obtain the total evaluation budget based on the preliminary prediction and future development trend.

[0069] The working principle and effects of the above technical solution are as follows: User interface construction and data entry: Using Vue.js to build the user interface, a form is created to collect user-entered historical R&D expense breakdown data. At this stage, the interface provides appropriate input fields for users to enter historical expense data for different projects. This data may include information such as project name, cost, year, and keywords. When the user submits the form, the system performs format validation to ensure that all input data conforms to predefined formats, such as costs must be numeric and years must be valid years. Preliminary R&D expense data forecast: Based on the historical R&D expense breakdown data provided by the user, the system uses pre-defined models or algorithms (such as linear regression and time series analysis) to generate preliminary expense forecasts. The purpose of this step is to predict possible future expenditures based on past data trends. Project development trend forecast: The system analyzes the keywords entered by the user for each project and combines them with external sources (such as industry reports, news articles, research papers, etc.) or internal historical data analysis to predict the future development trend of each project. This involves natural language processing technology to understand and analyze the correlation between keywords and market trends. Future trend factor: Based on the predicted project development trend, each project is assigned a future trend factor to quantify its development potential. This factor may be determined based on aspects such as the project's expected market growth rate and technological innovation potential; the preliminary forecast of R&D expenses and future trend factors are taken into consideration, and the final budget is calculated by weighted addition. The weighting here may take into account factors such as the importance, urgency and potential returns of the project to ensure the effective allocation of resources. It is highly accurate and combines the specific segmentation of historical data with forward-looking market trend analysis to more accurately predict future R&D expenses. This method is more scientific and reliable than predictions based solely on historical data or intuition; optimal resource allocation, through analysis of future trends, can help companies allocate resources more efficiently to the most promising projects, thereby increasing the return on R&D investment and the company's competitiveness;

[0070] Dynamic adaptability: The system design takes into account dynamic input and real-time analysis, enabling constant updates of data and budget forecasts, helping companies quickly adapt to market changes. The Vue.js-based front-end framework offers a smooth and intuitive user experience, a user-friendly interface, and clear data presentation, reducing operational complexity and improving data entry efficiency and accuracy. This technical solution, through the integrated use of front-end technology, data analysis, and forecasting models, provides companies with a scientific, comprehensive, and efficient R&D budget assessment tool. Through precise data entry verification, trend forecasting, and budget calculation, companies can optimize R&D resource allocation based on actual data and market trends, better plan future investments, enhance market competitiveness, and promote long-term development.

[0071] One embodiment of the present invention uses Vue.js to build a user interface and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense data, including:

[0072] Create a data entry form using HTML and CSS, and build the user interface using Vue.js in JavaScript. The interface construction code example is as follows:

[0073] <!DOCTYPE html>

[0074]

[0075]

[0076] <meta charset=""UTF-8"">

[0077] <meta name=""viewport"" content=""width=device-width,initial-scale=1.0"">

[0078] <title> R&D Expense Entry< / title>

[0079] <link rel=""stylesheet"" href=""styles.css"">

[0080]

[0081]

[0082]

[0083] <form @submit.prevent=""handleSubmit"">

[0084] <div v-for=""(entry," index) in entries" :key=""index"">

[0085] <label> Project Name:< / label>

[0086] <input type=""text"" v-model=""entry.name"required">

[0087] <label> cost:< / label>

[0088] <input type=""number"" v-model=""entry.cost"required">

[0089] <label> years:< / label>

[0090] <input type=""number"" v-model=""entry.year"required">

[0091]

[0092] <button type=""button"" @click=""addEntry""> Add an entry< / button>

[0093] <button type=""submit""> submit< / button>

[0094] < / form>

[0095]

[0096] <script src=""https: / / cdn.jsdelivr.net / npm / vue@2"">< / script>

[0097] <script src=""https: / / cdn.jsdelivr.net / npm / yup"">< / script>

[0098] <script src=""app.js"">< / script>

[0099]

[0100]

[0101] The JavaScript library Yup is used on the front end to perform preliminary verification to ensure that the user-entered data meets the requirements. The data verification and submission code examples are as follows:

[0102] new Vue({

[0103] el: '#app',

[0104] data: {

[0105] entries: [{ name: '', cost: '', year: ''}]

[0106] },

[0107] methods: {

[0108] addEntry() {

[0109] this.entries.push({ name: '', cost: '', year:''});

[0110] },

[0111] async handleSubmit() {

[0112] const schema = yup.array().of(

[0113] yup.object().shape({

[0114] name: yup.string().required(),

[0115] cost: yup.number().required(),

[0116] year: yup.number().required()

[0117] }) );

[0119] try {

[0120] await schema.validate(this.entries);

[0121] const response = await fetch(' / api / submit', {

[0122] method: 'POST',

[0123] headers: {

[0124] 'Content-Type': 'application / json'

[0125] },

[0126] body: JSON.stringify(this.entries)

[0127] });

[0128] const result = await response.json();

[0129] console.log('Submission successful', result);

[0130] } catch (error) {

[0131] console.error('Verification failed', error);

[0132] }

[0133] }

[0134] }

[0135] });

[0136] Submit the form data to the backend through the Fetch API;

[0137] The backend service receives the data and further verifies and processes the R&D data. The code example for further verification and processing is as follows:

[0138] const express = require('express');

[0139] const app = express();

[0140] app.use(express.json());

[0141] app.post(' / api / submit', (req, res) => {

[0142] const data = req.body;

[0143] / / Data validation and processing

[0144] / / Handle missing data and outliers

[0145] / / ...

[0146] res.json({ message: 'Data received successfully'});

[0147] });

[0148] app.listen(3000, () => {

[0149] console.log('The server is running at http: / / localhost:3000');

[0150] In one embodiment of the present invention, obtaining preliminary forecasted R&D expense data based on the historical R&D expense data includes:

[0151] Obtain R&D expense data from the past five years and process missing data and outliers in the R&D expense data. Use the data from the past five years to train an ARIMA model and select the optimal p and q values. The ARIMA model is:

[0152] Y t =c+φ1Y t-1 +φ2 Y t-2 +φ p Y t-p +…+θ1ε t-1 +…+θ q ε t-q +ε t

[0153] Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation;

[0154] Use the trained ARIMA model to predict R&D expenses;

[0155] Based on the output of the model, the prediction results are displayed through the Node.js visualization interface.

[0156] The effects of the above technical solution are: data centralization and standardization, through the web platform, the company's R&D expense data is centrally managed and verified to ensure the standardization and validity of the data; scientific prediction, using the ARIMA model to analyze and predict the R&D expense data of the past five years to provide scientific budget assessment; real-time visualization, using Node.js to visualize the prediction results, providing intuitive budget assessment information to facilitate decision-making; user-friendly, based on the Vue.js user interface design, providing a smooth user experience and reducing the difficulty of operation.

[0157] In one embodiment of the present invention, selecting optimal p and q values includes:

[0158] Draw a line graph of the historical R&D expense data obtained;

[0159] The stationarity of the time series is tested by the ADF test method, and the non-stationary time series is subjected to difference processing;

[0160] from statsmodels.tsa.stattools import adfuller

[0161] # ADF test

[0162] result = adfuller(data['expenses'])

[0163] print('ADF Statistic:', result[0])

[0164] print('p-value:', result[1])

[0165] # If p-value > 0.05, the data is non-stationary and requires differential processing

[0166] if result[1] > 0.05:

[0167] data_diff = data.diff().dropna()

[0168] result_diff = adfuller(data_diff['expenses'])

[0169] print('ADF Statistic after differencing:', result_diff[0])

[0170] print('p-value after differencing:', result_diff[1])

[0171] Draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) to preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part.

[0172] Use the AIC and BIC indicators to select the optimal combination of p and q, traverse different p and q values, compare the AIC and BIC values of each model, and select the combination that minimizes AIC or BIC to determine the optimal p and q values.

[0173] Use the statsmodels library to perform an ADF test to determine the stationarity of the time series. If the data is not stationary, perform a difference analysis. The sample code is:

[0174] from statsmodels.tsa.stattools import adfuller

[0175] # ADF test

[0176] result = adfuller(data['expenses'])

[0177] print('ADF Statistic:', result[0])

[0178] print('p-value:', result[1])

[0179] # If p-value > 0.05, the data is non-stationary and requires differential processing

[0180] if result[1] > 0.05:

[0181] data_diff = data.diff().dropna()

[0182] result_diff = adfuller(data_diff['expenses'])

[0183] print('ADF Statistic after differencing:', result_diff[0])

[0184] print('p-value after differencing:', result_diff[1])

[0185] Use AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion) to select the optimal combination of p and q. Traverse different p and q values, compare the AIC and BIC values of each model, and select the combination that minimizes AIC or BIC as the optimal p and q value. The sample code is:

[0186] import itertools

[0187] from statsmodels.tsa.arima.model import ARIMA

[0188] # Define the value range of p, d, q

[0189] p = q = range(0, 3)

[0190] d = [0, 1] # Usually d is 0 or 1

[0191] pdq = list(itertools.product(p, d, q))

[0192] # Traverse all combinations and select the model with the smallest AIC and BIC

[0193] aic_bic_values = []

[0194] for param in pdq:

[0195] try:

[0196] model = ARIMA(data['expenses'], order=param)

[0197] results = model.fit()

[0198] aic_bic_values.append((param, results.aic, results.bic))

[0199] except:

[0200] continue# Find the combination with the smallest AIC and BIC

[0201] best_aic = min(aic_bic_values, key=lambda x: x[1])

[0202] best_bic = min(aic_bic_values, key=lambda x: x[2])

[0203] print('Best AIC combination:', best_aic)

[0204] print('Best BIC combination:', best_bic)

[0205] In one embodiment of the present invention, a total evaluation budget is obtained by weighted addition of future development trends and preliminary predictions, including:

[0206] Obtain the evaluation project categories, design a web crawler using Python's Scrapy library, and crawl articles and post data on the forum;

[0207] Organize the collected data by time and count the number of articles and posts in each month;

[0208] Categorize article and post content to identify content relevant to the assessment project category;

[0209] Calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category;

[0210] Use the weighted moving average method to smooth the growth rate, and then calculate the trend score of the project category based on the weighted moving average growth rate;

[0211] The total annual R&D expenditure of the assessed enterprise is obtained based on the trend score and the project expenditure predicted by ARIMA.

[0212] Calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category;

[0213]

[0214] in, is the growth rate in time period t, is the number of articles and posts in time period t, is the number of articles and posts in time period t-1;

[0215] Use the weighted moving average method to smooth the growth rate, and then calculate the trend score of the project category based on the weighted moving average growth rate;

[0216] WMAt =

[0217] Among them, WMA t is the weighted moving average growth rate in time period t, is the weight coefficient, usually the most recent data has a greater weight, and k is the time window size of the moving average.

[0218] S=

[0219] in, is the weighted moving average growth rate of the last time period, is the weighted moving average growth rate of the first time period, is the number of time periods, S is the trend score; the total annual R&D expenses of the assessed enterprise are obtained based on the trend score and the project expenses predicted by ARIMA.

[0220] St= Sj×(1+S)

[0221] Among them, St is the total evaluation cost, Sj is the predicted cost of each project, S is the trend score of the project, and m is the total number of R&D projects of the enterprise.

[0222] One embodiment of the present invention provides a web-based platform-based enterprise annual R&D expense budget assessment system, the system comprising:

[0223] The data acquisition module is used to build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense data;

[0224] A preliminary forecasting module, configured to obtain preliminary forecasted R&D expense data based on the historical R&D expense breakdown data;

[0225] Development trend scoring module, used to obtain a total evaluation budget based on preliminary forecasts and future development trends

[0226] The total evaluation module is used to obtain the total evaluation budget based on the weighted addition of future development trends and preliminary forecasts.

[0227] In one embodiment of the present invention, a system for evaluating an enterprise's annual R&D expense budget based on a web platform is provided, wherein the data acquisition module includes:

[0228] Build interface modules for creating data entry forms with HTML and CSS, and building user interfaces using Vue.js in JavaScript;

[0229] The validation format module is used to perform preliminary validation on the front end using the JavaScript library Yup to ensure that the data entered by the user meets the requirements;

[0230] Submit backend module, used to submit form data to the backend through Fetch API;

[0231] The processing module is used for the back-end service to receive data and further verify and process the R&D data.

[0232] In one embodiment of the present invention, a web-based platform-based enterprise annual R&D expense budget assessment system is provided, wherein the preliminary forecast module includes:

[0233] A historical data acquisition module is used to obtain R&D expense data for the past five years and process missing data and outliers in the R&D expense data;

[0234] The training model module is used to train the ARIMA model using the data from the past five years and select the optimal p and q values. The ARIMA model is:

[0235] Y t =c+φ1Y t-1 +φ2 Y t-2 +φ p Y t-p +…+θ1ε t-1 +…+θ q ε t-q +ε t

[0236] Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation;

[0237] The forecasting module is used to forecast R&D expenses using the trained ARIMA model;

[0238] The visualization display module is used to display the prediction results through the Node.js visualization interface based on the output of the model.

[0239] In one embodiment of the present invention, a system for evaluating an enterprise's annual R&D expense budget based on a web platform is provided, wherein the training model module includes:

[0240] A graph drawing module is used to draw a line graph of the acquired historical R&D expense data;

[0241] The stationary detection module is used to detect the stationarity of the time series through the ADF test method and perform differential processing on non-stationary time series;

[0242] A preliminary determination module is used to draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) graphs, and preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part.

[0243] The module for determining the optimal p and q values is used to select the optimal p and q combination using the AIC and BIC indicators. It traverses different p and q values, compares the AIC and BIC values of each model, and selects the combination that minimizes AIC or BIC to determine the optimal p and q values.

[0244] One embodiment of the present invention provides an enterprise annual R&D expense budget evaluation system based on a web platform, wherein the overall evaluation module includes:

[0245] The crawling module is used to obtain the evaluation project categories and design a web crawler using Python's Scrapy library to crawl articles and post data on the forum;

[0246] The statistics module is used to organize the collected data by time and count the number of articles and posts in each month;

[0247] Identification module, used to classify the content of articles and posts and identify content related to the assessment item category;

[0248] The growth rate calculation module is used to calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category;

[0249] The trend score calculation module is used to smooth the growth rate using the weighted moving average method, and then calculate the trend score of the project category based on the weighted moving average growth rate;

[0250] The total evaluation module is used to obtain the total annual R&D expenses of the evaluated enterprise based on the trend score and the project expenses predicted by ARIMA.

[0251] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for evaluating an enterprise's annual R&D expense budget based on a webpage platform, characterized in that: The method comprises: Build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data; Obtain preliminary forecast R&D expense data based on the historical R&D expense breakdown data, including: Obtain R&D expense data for the past five years and address missing data and outliers in the R&D expense data; Using the data from the past 5 years, we train an ARIMA model and select the best p and q values. The ARIMA model is: Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation; Use the trained ARIMA model to predict R&D expenses; Based on the output of the model, the prediction results are displayed through the Node.js visualization interface; Predict the future development trend of the project based on the project keywords, and obtain the total evaluation budget based on the preliminary prediction and future development trend, including: Obtain the evaluation project categories, design a web crawler using Python's Scrapy library, and crawl articles and post data on the forum; Organize the collected data by time and count the number of articles and posts in each month; Categorize article and post content to identify content relevant to the assessment project category; Calculate the growth rate of the number of articles and posts in each month to reflect the changes in the popularity of the project category: in, is the growth rate in time period t, is the number of articles and posts in time period t, is the number of articles and posts in time period t-1; Use the weighted moving average method to smooth the growth rate, and then calculate the trend score of the project category based on the weighted moving average growth rate: Among them, WMA t is the weighted moving average growth rate for time period t, is the weight coefficient, usually the most recent data has a greater weight, and k is the time window size of the moving average; in, is the weighted moving average growth rate of the last time period, is the weighted moving average growth rate of the first time period, is the number of time periods, S is the trend score; The total estimated annual R&D expenditure of the enterprise is obtained based on the trend score and the project expenditure predicted by ARIMA: Among them, S t is the total assessment fee, S j is the predicted cost of each project, S is the trend score of the project, and m is the total number of R&D projects of the enterprise.

2. The method for evaluating an enterprise's annual R&D expense budget based on a webpage platform according to claim 1, characterized in that: Build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data, including: Create data entry forms with HTML and CSS, and build user interfaces using Vue.js in JavaScript; Use the JavaScript library Yup on the front end to perform preliminary verification to ensure that the data entered by the user meets the requirements; Submit the form data to the backend through the Fetch API; The backend service receives the data and further verifies and processes the historical R&D expense breakdown data rows.

3. The method for evaluating an enterprise's annual R&D expense budget based on a webpage platform according to claim 1, characterized in that: Select the best values of p and q, including: Draw a line graph of the historical R&D expense data obtained; The stationarity of the time series is tested by the ADF test method, and the non-stationary time series is subjected to difference processing; Draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) to preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part. Use the AIC and BIC indicators to select the optimal combination of p and q, traverse different p and q values, compare the AIC and BIC values of each model, and select the combination that minimizes AIC or BIC to determine the optimal p and q values.

4. A web-based enterprise annual R&D expense budget assessment system, characterized by: The system comprises: A data acquisition module is used to build a user interface using Vue.js and create a data entry form to obtain the user's historical R&D expense breakdown data and verify the format of the historical R&D expense breakdown data; A preliminary forecasting module is used to obtain preliminary forecast R&D expense data based on the historical R&D expense breakdown data, including: A historical data acquisition module is used to obtain R&D expense data for the past five years and process missing data and outliers in the R&D expense data; The training model module is used to train the ARIMA model using the data from the past five years and select the optimal p and q values. The ARIMA model is: Among them, Y t represents the R&D expenditure at time t, c is a constant term, representing the average level of R&D expenditure, φ p represents the impact of R&D expenses at p time points in the past on current R&D expenses, θ q represents the impact of the error term at the past q time points on the current R&D expenses, ε t represents the random error term at time t, which represents the current unpredictable random fluctuation; The forecasting module is used to forecast R&D expenses using the trained ARIMA model; The visualization display module is used to display the prediction results through the Node.js visualization interface based on the output of the model; The total evaluation module is used to predict the future development trend of the project based on the project keywords and obtain the total evaluation budget based on the preliminary prediction and future development trend, including: The crawling module is used to obtain the evaluation project categories and design a web crawler using Python's Scrapy library to crawl articles and post data on the forum; The statistics module is used to organize the collected data by time and count the number of articles and posts in each month; Identification module, used to classify the content of articles and posts and identify content related to the assessment item category; The growth rate calculation module is used to calculate the growth rate of the number of articles and posts in each month to reflect the popularity changes of the project category: in, is the growth rate in time period t, is the number of articles and posts in time period t, is the number of articles and posts in time period t-1; The trend score calculation module is used to smooth the growth rate using the weighted moving average method, and then calculate the trend score of the project category based on the weighted moving average growth rate: Among them, WMA t is the weighted moving average growth rate in time period t, is the weight coefficient, usually the most recent data has a greater weight, and k is the time window size of the moving average; in, is the weighted moving average growth rate of the last time period, is the weighted moving average growth rate of the first time period, is the number of time periods, S is the trend score; The evaluation module is used to obtain the total estimated annual R&D expenditure of the enterprise based on the trend score and the project expenditure predicted by ARIMA: Among them, S t is the total assessment fee, S j is the predicted cost of each project, S is the trend score of the project, and m is the total number of R&D projects of the enterprise.

5. According to claim 4, a web platform-based enterprise annual R&D expense budget evaluation system is characterized in that: The data acquisition module includes: Build interface modules for creating data entry forms with HTML and CSS, and building user interfaces using Vue.js in JavaScript; The validation format module is used to perform preliminary validation on the front end using the JavaScript library Yup to ensure that the data entered by the user meets the requirements; Submit backend module, used to submit form data to the backend through Fetch API; The processing module is used for the back-end service to receive data and perform one-step verification and processing on the historical R&D expense breakdown data.

6. According to claim 4, a web platform-based enterprise annual R&D expense budget evaluation system is characterized in that: The training model module includes: A graph drawing module is used to draw a line graph of the acquired historical R&D expense data; The stationary detection module is used to detect the stationarity of the time series through the ADF test method and perform differential processing on non-stationary time series; A preliminary determination module is used to draw the autocorrelation function (ACF) and partial autocorrelation function (PACF) graphs, and preliminarily determine the p and q values. ACF is used to determine the order q of the MA part, and PACF is used to determine the order p of the AR part. The module for determining the optimal p and q values is used to select the optimal p and q combination using the AIC and BIC indicators. It traverses different p and q values, compares the AIC and BIC values of each model, and selects the combination that minimizes AIC or BIC to determine the optimal p and q values.

Images